1IIC AI Report 2020

Artificial Intelligence
in the Asia-Pacific
Region
Examining policies and strategies to
maximise AI readiness and adoption
February 2020
2 International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region
Contents
I. Executive Summary: Creating and Seizing AI Opportunities 6

The Promise of AI 7
The Challenge of AI 9
The Policy Response to AI 12
II. The Promise of AI: Solving Problems, Improving Lives 14

What AI Already Solves 16
What AI Will Improve 18
III. The Challenge of AI: Key Topics to Address 20

Key Topic 1: Infrastructure 22
Key Topic 2: Access to Data 24
Key Topic 3: Skills and Human Capital 27
Key Topic 4: Trust and Partnerships 29
Key Topic 5: Ecosystem and Entrepreneurship 32
IV. The Policy Response to AI: Overview of AI Strategies 34

Global AI Policy Strategies 35
APAC AI Policy Strategies 40
V. Conclusion: Design Principles for Effective AI Policies 75

Design Principle 1: Shaping AI Ecosystems 76
Design Principle 2: Preparing AI Talent 78
Design Principle 3: Building Trust in AI 79
VI. References 81
International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region 3

Introduction
The International Institute of Communications (IIC) This document is designed to enable stakeholders
has continually responded to the exponential growth to consider different policy options as they address
in the communications and media sector, with the issues raised by AI, so that they might build a
digitally-driven disruption rife throughout the various framework that enables its development and use as
layers that make up this grouping of hardware, a positive force in their countries, while mitigating
networks, services, and content. It recognised risks and allaying concerns. Commissioned from the
too, the underpinning role of Artificial Intelligence consultancy, TRPC, based in Singapore, it draws on
(AI), driving convergence not only across our personal interviews and desk research, examining
telecommunications, media and technology sectors, current policies in the Asia-Pacific region, but also
but blurring the boundaries between many industry ventures farther afield and looks at the way in which
sectors. emerging policies are developing.
The IIC has always addressed policy issues head

on, with our first dedicated meeting to the subject
of AI in Mexico (2018). An overview of the issues
raised by AI has been addressed in a briefing paper
published in 2019 and during all subsequent IIC
events, as well as in regular publications, such as its Andrea Millwood Hargrave
journal Intermedia. Director General, International Institute of Communications

About the IIC
The IIC has operated at the intersection of business and
regulation for over 50 years. By convening innovators
and policy makers the IIC helps find regulatory
frameworks to protect the interests of individuals,
commercial organisations, economies and wider
society. IIC members share a common belief that the
impact of technological innovation on society should be
evaluated and discussed and that policy and regulation
should not hinder innovation and investment.
Learn more at www.iicom.org
About TRPC
TRPC is a boutique consulting and research firm with
over 25 years’ experience in the telecommunications
and ICT industries in the Asia-Pacific. We offer
specialised advisory, research, and training services,
with a focus on regulatory and strategic business issues,
and possess an extensive network of industry experts
and professionals throughout the region.
For more information, please visit www.trpc.biz

I.
Executive
Summary

Creating and Seizing AI Opportunities
A fast-evolving technology that can serve multiple purposes, artificial

intelligence (AI) does not have a single, universally accepted definition.
For the purposes of this report, “AI” designates all The Promise of AI
computer systems that can reason, learn, and act
in response to what they sense and perceive. This Though still in their infancy, AI products and services
encompasses everything from machine learning are being introduced across finance, healthcare,
programmes and intelligent algorithms to predictive retail, transport, manufacturing, and agriculture –
analytics, autonomous devices, and automated strategic industries that together drive the economic
systems. Box 1 at the end of this section presents dynamism of both emerging and mature APAC
the different levels and types of AI currently in economies.1
existence.
For example, several major Asian insurers such as
The different facets of AI present a range of Prudential Singapore, Ping An, and Sompo Japan
economic opportunities while bringing complex use machine learning – systems that learn and
regulatory challenges to the forefront. The scope adjust their behaviour based on the information
and pace of technological advancements made by they receive – to automatically assess claims and
AI are pushing decision-makers and policy leaders manage fraud.2 In China, Alibaba deployed its
in the Asia-Pacific (APAC) region – as elsewhere – proprietary ET Agricultural Brain initiative, an AI-
to take a closer look at the way it is developed and based project that uses visual recognition and real-
implemented. time environmental monitoring to track the growth
conditions of crops and livestock. Through machine
The following offers an overview of developments learning algorithms, the data gathered helps farmers
in AI and the way it is being implemented and cut costs and increase yields.3
promoted across APAC economies.

According to PwC estimates, the broad field of AI In India, Accenture is working with the National
will contribute up to USD15.7 trillion to the global Association for the Blind to deploy an AI-powered
economy in 2030 – more than the current output of system that uses image recognition, natural
China and India combined. Of this, USD6.6 trillion language processing, and natural language
are foreseen to come from increased productivity generation to inform visually challenged users on
and USD9.1 trillion from consumption-side effects.4 In the number of people in a room, their ages, genders,
this context, AI is rapidly becoming central to public- and even emotions based on facial expressions.
and private-sector organisations’ strategies across The system can also be used to identify and narrate
the region. text from books and documents, including currency
notes, and identify obstructions.14
For businesses, AI provides immediate potential
productivity gains – enabling the automation
of tasks, the streamlining of processes, and the
optimisation of resources – and can create long-
term competitiveness by boosting overall investment
and innovation. According to McKinsey estimates,
for example, multinational corporations such as In this sense, policy-makers can
Alibaba, Amazon, Google, Baidu, and Facebook determine the extent to which
invested between USD20 billion and USD30 billion
AI becomes a tool for societal
globally in AI in 2016.5 For governments, AI can
help address a range of issues, including many
good. From best practices to
of the United Nations’ Sustainable Development ethical principles, policy-makers
Goals (SDGs). From poverty reduction to natural can help prepare individuals,
disaster mitigation, AI used as a full-fledged national organisations, and communities to
priority and not just as a cost-cutting measure can take advantage of the benefits of
be leveraged to tackle complex and longstanding AI and its associated innovations,
challenges.6
while mitigating any unintended or
as-yet unidentified risks.
To this end, the governments of China,7 Singapore,8
and South Korea9 are investing heavily in building AI
ecosystems in which AI-driven companies and their
innovative business models can grow and mature.
In Indonesia,10 Malaysia,11 and Thailand,12 numerous
public-private initiatives have begun taking shape
to support experimental AI projects that go beyond
for-profit uses of AI.
Other examples include the engineering student

in the Philippines who has leveraged AI to
communicate with patients suffering from locked-
in syndrome – a severe paralysis due to traumatic
brain injury, neural damage, or stroke that results
in total loss of motor ability, including speech. The
project involves a brain-computer typing interface
that records brain signals and uses machine learning
to turn them into basic shapes and letters.13

The Challenge of AI
The growth of AI across sectors, societies, and economies brings

complex, inter-connected challenges to the forefront for national
agenda-setting; and these need to be considered carefully – but
promptly – as we move forward into an era of digital transformation.
The challenges can be bucketed together under four AI ‘topics’ that are
at the forefront of APAC policy-makers’ discussions as they define and
frame national approaches to AI:
Infrastructure
Strong infrastructural foundations that provide reliable, widespread, and affordable

connectivity, are a fundamental prerequisite in allowing organisations (both commercial
and non-commercial) and individuals to access digital platforms.
Even when a population understands AI and welcomes the changes that can be
engendered, it does not necessarily translate into access, usage, and development
if there is limited reach. This is especially true for rural and remote areas of the APAC
region, where poor infrastructure and even poorer connectivity hinder populations’ ability
to participate in the digital economy.15
According to the ITU, close to half of the world’s population still does not have access to
the internet, with Asia and Africa having the lowest rates of access.16 Across APAC, close
to 417 million people do not have access to basic internet services.17 At the same time,
only 16% of the region’s SMEs use digital tools, with half citing price as a major obstacle.18

Access to data
Globalisation and digitalisation have led to greater connectivity, which in turn has rapidly
increased the quantities of data flowing between devices, platforms, and networks. In
this context, the ability to access and transfer data across boundaries and jurisdictions
is increasingly crucial to the data-driven technologies that consumers, businesses, and
governments use on a daily basis – including AI.
Yet many APAC economies have regulatory measures in place that limit or restrain the
flow of data. Designed to keep data secure by confining it within a country’s borders,
data localisation measures can take many shapes; they can be explicitly required by law
(within a cybersecurity regulation, for instance) or result from a combination of policies
that makes it expensive, complicated, or unfeasible to transfer data transnationally.
In some cases, such barriers target the protection of consumers’ private information,
limiting any form of unauthorised data transfer. In other cases, governments use them
to ensure sensitive or strategic data cannot be accessed or compromised by foreign
entities.
Whichever the reason, such measures can negatively impact the emergence and
maturation of AI. AI needs vast amounts of data to effectively learn, grow, and expand,
and blanket restrictions on data movements hinder its ability to access this vital
resource. A key challenge for governments across the region is the design of policies
and regulations that strike a balance between addressing valid privacy and security
concerns and enabling the movement of data to keep their economies dynamic,
competitive, and relevant in the digital age.
Skills and human capital
A key concern for citizens and governments alike is the extent to which workers and
businesses are equipped with the skills, knowledge, and confidence to both use and
benefit from AI systems.
But the pace and scope of AI’s progress across sectors and industries can be difficult
to follow from an institutional and organisational point of view. In many APAC countries,
technological advances take place much faster than decision-makers can define
approaches to fully harness or leverage them.
For businesses, this is embodied by the challenge of adapting workers, processes,

and business models to radically reshaped competitive dynamics.19 For governments,
the process can make it difficult to design policies that balance the need to support AI
ecosystems (developing an AI-capable workforce) with the obligation to shield citizens
from harmful unintended consequences (upholding workers’ rights and interests).20
This picture suggests the need for a new, more flexible and cooperative policy-framing
approach.

Trust and partnerships
Developing data-driven innovations such as AI in a safe, ethical, and sustainable manner

is necessary to enable consumers’ trust and understanding, and to help get citizens’
“buy-in” to the AI development trajectory. A complex and constantly evolving field,
AI encompasses different types of systems (machine learning, deep learning, neural
networks, autonomous vehicles, etc.) and can serve a range of purposes in a broad
spectrum of sectors and industries. For example, the AI used by financial institutions is
often not the same as the AI deployed in hospitals or factories.
This variety of hard-to-define technologies and systems can sometimes make it

challenging to understand AI and where the respective systems will develop. Surveys
show that despite AI awareness, citizens, consumers, business leaders, and even public-
sector decision-makers do not necessarily understand the way AI technologies work –
leading to a reluctance to trust and embrace them.21
Another contributing factor is the rising trend of businesses rebranding existing products
and services as “AI” to attract capital and customers. From data analytics to predictive
modelling, data-mining systems have long been used to process large amounts of data –
only today many of them are referred to as “AI”. This muddles the general understanding
of AI and exacerbates both inflated expectations and misplaced suspicions. Providing
clarity and developing signposts can serve to assure and allay many of the more
challenging and far-reaching concerns, and contribute to a more constructive and
coordinated approach to pursuing AI for the betterment of society.
Ecosystem and entrepreneurship
The emergence of a local AI industry – and a local AI start-up scene – is vital for
economies aiming to become regional and global AI leaders.
At the national level, APAC governments are investing heavily in developing their own
homegrown AI capabilities by financing research, development, and deployment efforts.
Many, such as Australia, New Zealand, and South Korea, do so through coordinated local
and national strategies, while others, such as Indonesia, Malaysia, and Thailand, adopt a
more collaborative approach with private-sector players. Others still, such as China and
Singapore, do both.
It is worth noting that China and Singapore are already looking further into the future,
building strong foundations to become regional and global AI hubs. Indeed, their national
AI strategies go one step further than creating dynamic and conducive environments
in which innovative and competitive AI businesses can emerge; they explicitly aim to
position their environments as attractive platforms for regional and global AI ecosystems
to grow and mature.
In this review we show how these four topics are reflected in the policy and regulatory responses that
can be observed around the APAC region, albeit with varying priorities and focus areas.

The Policy Response to AI respects human rights and inclusively shares the
benefits and opportunities it creates.24
The AI policy landscape is both nascent and diverse These various approaches to AI policy-making have
in the seven APAC economies covered in this report. at least one element in common; they demonstrate
Nascent because there are few AI-specific policies that designing policies for AI requires much
driving AI adoption and readiness on a national more than addressing technological or economic
level. Where there are AI policies in place, they are challenges.
mostly sector-specific plans to upgrade strategic
industries with AI or to prepare a sector’s workforce This report examines the different AI policies and
to AI-intensive environments. Diverse because strategies governments in the region are beginning
countries’ approaches to AI vary greatly based on to introduce, the key issues they are looking at,
two distinct but inter-related elements: the different and the specific challenges they are attempting to
priorities that have been defined and the different address. It also provides in-depth country studies
resources that have been mobilised to harness the for seven APAC economies, chosen for the variety
economic, social, ethical, and legal implications of AI of AI-specific strategies they are designing and
advancement. adopting. The report highlights examples of public-
and private-sector initiatives that are driving national
For example, China and Singapore have broad, AI readiness and adoption, and cites a number of
overarching national policies specifically designed AI insights collected from interviews with AI experts
to drive AI readiness and to regulate its deployment, (researchers, academics, regulators, and industry
whereas Malaysia and South Korea have AI-related players).
actions within broader digital transformation plans
(though they both have announced forthcoming
national frameworks aimed at consolidating the
wide range of siloed initiatives in place).
Box 1.
Indonesia and Thailand, meanwhile, have launched Levels and Types of Artificial Intelligence
targeted initiatives and programmes that facilitate
the adoption of AI systems in strategic sectors and There is no single, universally accepted
industries, focusing on public-private cooperation definition of AI. A complex and
to drive adoption. Lastly, Australia has no formal constantly evolving field, there are as
AI policy in place, but has a variety of official many possible definitions as there are
guidelines, principles, and standards to help potential uses and applications.
individuals, businesses, and institutions prepare for
AI-driven disruption. And this confusion will only grow as
increasingly complex AI innovations
Alongside these national approaches are come out of research laboratories and
noteworthy multi-lateral strategies. The OECD, make their way into mainstream society,
for instance, is putting AI at the top of members’ transforming the way humans and
policy agendas via its Principles on Artificial machines live and work together.
Intelligence, the first international standards agreed
by governments for the responsible stewardship Despite this difficulty, it is possible to
of trustworthy AI.22 It has also created the OECD distinguish between several broad types
AI Policy Observatory, an online hub providing and levels of AI. The figures below
resources on AI policies, initiatives, and metrics.23 present some of the most common
forms of AI, based on function, purpose,
Likewise, the G20 has formulated its own AI and complexity.
Principles, highlighting the importance of a human-
centred approach to AI – using AI in a way that

Levels Of Artificial Intelligence
Types of Artificial Intelligence
Source: Deloitte, www2.deloitte.com/fi/fi/pages/technology/articles/part1-artificial-intelligence-defined.html;

PwC, www.pwc.com/gx/en/issues/data-and-analytics/publications/artificial-intelligence-study.html;
Intellectyx, www.intellectyx.com/artificial-intelligence

II.
The Promise
of AI

Solving Problems, Improving Lives
With its young, tech-savvy population and a marked appetite for innovation,
the APAC region has what it needs to adopt and deploy increasingly smart,
hyper-connected technologies that power digitally enabled societies.
This is also true for AI, which, as the AI for Good

Global Summit notes, can be used to address some
of the greatest challenges facing humanity.25
In APAC countries, many initiatives are already

leveraging AI to fulfil a number of the United
Nations’ Sustainable Development Goals (SDGs)
– hunger, poverty, inequality, discrimination. And
just as many experimental projects are working on
mitigating the impact of more complex issues –
climate change, water security, air quality, to name
but a few.

What AI Already Solves
The widespread digitisation of human activity generates the kind of

massive data sets necessary for AI applications and systems to perform
highly complex functions. Many such systems are already used to
automate and optimise tasks that would require a less efficient use
of manpower and resources to complete. And they are doing so in a
wide range of areas that goes well beyond the most common industrial
applications of AI. Some examples follow:
AI tracks mosquito populations to reduce the spread of dengue

(Singapore)
Project Wolbachia – a joint project between the National Environment

Agency (NEA) and Verily (formerly Google Life Sciences) – is a multi-
year project that aims to curb the dengue virus-carrying Aedes
aegypti mosquitoes. The idea is to inject male mosquitoes with the
Wolbachia bacteria, which is transmitted to females during mating and
prevents the hatching of mosquito eggs. Using an AI-based sex-sorting
technology to accurately separate the males from the females, the NEA
ensures the sorting is less laborious and more accurate than a more
manual approach.26
AI-powered microscope automatically spots Malaria parasites

(China)
The EasyScan GO is an AI-enabled microscope developed to assist

clinicians in identifying malaria parasites in blood smears. Created
by Chinese microscope manufacturer Motic China Group, EasyScan
GO accurately detects severe and drug-resistant cases of malaria,
bypassing the need for blood-smear analysis by a World Health
Organisation (WHO)-certified expert microscopist. This automated
process provides results faster than with a manual examination,
and alleviates the shortfall of trained personnel in under-resourced
countries.27
AI programme automatically screens for diabetic eye disease

(Thailand)
Google is working with the state-run Rajavithi Hospital to launch an

AI programme that analyses patients’ eye-screen results to assess if
they are at a risk of diabetic eye disease. The programme analyses
patients’ eye screening results to assess if they are at a risk of partial
or complete vision loss, prompting the provision of vital pre-emptive
treatment. The system has an accuracy rate of 95% when it comes to
disease detection, compared with 74% from opticians or eye doctors.28

AI system allows the visually impaired to live more autonomously
(India)
Accenture is working with the National Association for the Blind in India
to deploy Drishti, an AI-powered system that uses image recognition,
natural language processing, and natural language generation to provide
contextual information via users’ smartphones. The system can describe
the number of people in a room, their ages, genders, and even emotions
based on facial expressions. It can also be used to identify and narrate
text from books and documents, including currency notes, and identify
obstructions like glass doors.32
AI enhances security in urban centres (Malaysia)
Since 2017, MIMOS, Malaysia’s national research and development

centre for ICT, has been working with Huawei to develop AI-based
security and surveillance solutions, including advanced video analytics
and facial recognition systems. The solutions help cities automate
everything from crime-detection mechanisms to emergency-response
processes, making public services faster and more efficient.33
AI module customises students’ learning experiences (Vietnam)
Vietnamese tech giant FPT developed VioEdu, an AI-powered online

e-learning system that allows students to personalise their learning
experiences. The module examines students’ behaviours during
training courses to model their individual learning patterns. Once
established, the patterns are synthesised and analysed by the system,
allowing it to provide contents and delivery methods in a tailored
manner. It can also identify potential knowledge gaps to propose
suitable learning pathways.29
AI improves customer service and enhances e-commerce (Indonesia)
Kata.ai, Indonesia’s most popular conversational AI platform, works with

Unilever and Telkomsel to streamline and automate millions of customer
interactions – up to 96% of customer queries – with minimum human
interaction.30 Tokopedia, Indonesia’s biggest internet marketplace and
one of the country’s five unicorns, launched an AI research centre
in conjunction with the University of Indonesia (UI) to develop AI-
based solutions for the e-commerce sector (demand prediction, smart
warehouses, and smart logistics).31

What AI Will Improve
Whether directly or indirectly, it is increasingly clear that no sector or

industry will go unaffected by AI. PwC notes that one area in which
AI can have the most visible – and meaningful – impact is in helping
mitigate environmental damages and risks.34 From the creation of
distributed “off-grid” water and energy resources to improved natural
disaster resilience planning, there is a wide range of areas in which AI
can effectively combat the unintended effects of industrial activities:
Distributed energy grids
AI can enhance the predictability of demand and supply for renewable

energy sources across distributed grids, improving energy storage and
efficiency. It can also support load management, assist in the integration
and reliability of renewables, and enable dynamic pricing and trading.
Smart agriculture and food systems
AI-augmented agriculture involves automated data collection and

data-based decision-making to allow corrective measures to be
made automatically and in real-time. Whether it is the detection of
crop diseases or the nutrition levels of livestock, AI can help increase
resource efficiencies across the agriculture industry, lowering the use of
water, fertilisers, and pesticides.
Weather and climate science and forecasting
A field that was once limited by its requirement of high-performance,

energy-intensive computing power, climate informatics now benefits
from deep-learning networks that can calculate millions of variables
simultaneously and faster than ever before.

Smart disaster response
AI can analyse simulations and real-time data to predict weather events

and disasters, allowing citizens and relevant authorities to take action
in a timely and proactive manner. Deep-learning capabilities can also
enhance disaster simulations to determine optimal response strategies,
similar to the way AI is currently being used to identify the best moves in
games like Go.
AI designed intelligent, connected, and liveable cities
AI could also be used to simulate and automate the generation of

zoning laws, building ordinances, and floodplains. Combined with
augmented and virtual reality (AR and VR) modules, AI can provide real-
time city-wide data on energy, water consumption, traffic flow, people’s
movements, and weather hazards – all of which could optimise urban
sustainability plans.
A transparent, digital Earth
A real-time, open-architecture, AI-infused digital geospatial dashboard

for the planet would enable the monitoring, modelling, and management
of environmental systems at a scale and speed never before possible
– from identifying illegal deforestation and water extraction to tackling
biodiversity loss and air pollution. The AI technologies exist, but what is
missing is collaboration for open source data and APIs on a global scale.

III.
The Challenge
of AI

Key Topics to Address
AI can become a truly positive force for societies, communities, and

economies in APAC. Five key topics are at the heart of APAC policy-makers’
discussions as they design frameworks to harness the power of AI, mitigate
any potential risks, and allay concerns:
1. Infrastructure – Strong infrastructural foundations provide reliable and widespread connectivity,

allowing individuals and organisations to access digital platforms. It is much more difficult for a
dynamic and innovative digital economy to emerge if only part of the population can contribute to it.
2. Access to data – Data and information are the key drivers of digital products and services.
From collection to processing and transfer, data that flows freely and securely across platforms,
borders, and jurisdictions can help data-driven AI technologies learn, grow, and expand.
3. Skills and human capital – Workers and businesses who are equipped with computer
science skills and knowledge can both contribute and benefit from the digital economy. In
the case of AI, being able to use AI systems will ensure no one is left behind as workplaces
and business models turn to automation.
4. Trust and partnerships – Developing AI in a safe, ethical, and sustainable manner can
increase citizens’ and consumers’ trust and understanding, which in turn improves the
pace and scope of AI adoption. Multi-stakeholder collaboration, meanwhile breaks
down traditional silos that may hinder the development and spread of AI.
5. Ecosystem and entrepreneurship – For economies to remain relevant and

competitive in the age of AI, it is important for a local AI industry – and a local AI
start-up scene – to emerge and mature. The right enabling regulatory and policy
frameworks can support the growth of a national AI value chain, thereby creating a
strong and sustainable AI ecosystem.

Key Topic 1: Infrastructure 16% of the region’s SMEs actually use digital tools,
and half of them cite price as a major obstacle.39
Strong infrastructural foundations provide the
reliable and widespread connectivity that AI – Affordability is indeed a major barrier: the price of
and many other digital technologies – require to mobile data has generally declined, but the fact
be effective and useful. This includes a reliable that applications use much more data than before
national broadband telecommunications network, makes internet use much too expensive for a large
a comprehensively regulated telecommunications part of the region’s population.40 Data from countries
sector, investments in improved and widespread holding 4G spectrum auctions between 2008 to
connectivity (especially in rural or remote areas), 2016 shows that profit-oriented spectrum pricing
and secure and reliable access to cloud computing may negatively impact consumers’ purchasing
systems. power as well as hinder foreign investment inflows.41
In this context, the regulation of the

telecommunications sector plays a role in improving
connectivity throughout the region.
Infrastructure for inclusive internet access
The APAC region is particularly impacted by a

widening digital divide. A region fraught with deep
socio-economic inequalities, these are reflected
in an uneven and fragmented access to digital
technologies.
The penetration of digital communications is at

an all-time high in the region. According to the
GSMA, 2.8 billion people in APAC subscribed to
Infrastructure for affordable internet access mobile services in 2018, and the region has been
the biggest contributor to global subscriber growth
Affordable and reliable connectivity to regional in recent years, with 466 million new subscribers
and global telecommunications networks is added since 2014.42
indispensable for the growth and development of AI.
And yet the digital divide persists. According to the
Some 370 million Southeast Asians had access World Bank, the share of the APAC population living
to the internet in 2018, thanks largely to the rapid in extreme poverty (living with USD1.90 or less per
expansion and the continued improvement of both day) has fallen below 3%, but that still leaves some
broadband and mobile infrastructure.35 Smartphones 47 million people in extreme poverty.43 These are 47
account for more than 90% of Southeast Asian million people for whom e-commerce, e-payments,
internet users, a trend that could accelerate intra- and even smartphone ownership are far from being
regional trade and growth, enabling a collective day-to-day priorities.
GDP uplift of USD1 trillion by 2025.36
Even within rapidly digitalising markets, connectivity
But despite these encouraging trends, poor is not evenly distributed across urban and rural
infrastructure and even poorer connectivity remain populations. In Indonesia, only about 42-48%
major issues for rural and remote areas.37 Close of people in rural communities have access to
to 417 million people – most of them in Cambodia, smartphones and the internet, compared to 71-72%
Laos, Myanmar, and Vietnam – do not have access of people in urban cities.44
to basic internet services.38 At the same time, only

The digital divide is further exacerbated by the many of the tasks done manually. The underlying
region’s unequal vulnerability to natural and man- architecture facilitates elastic consumption, self-
made catastrophes. Some regions experience service, and pay-as-you-go pricing, thereby
disasters much more frequently and intensely than providing substantial economies of scale.48 This
others, leading to the disruption of infrastructure that makes it possible to bypass complex or rigid
may not have been designed to withstand extreme technological requirements, making it easier
conditions. Earthquakes, for instance, frequently and more affordable to provide advanced digital
damage submarine cables and interrupt internet products and services.
access in densely populated coastal areas and
cities.45 Cited in a report, Michael Zielenziger, Managing
Editor of Oxford Economics, sees cloud computing
Dr Shamshad Akhtar, Executive Secretary of UN- as “(…) the ultimate ‘leapfrog’ technology. It allows
ESCAP, offers some insights into internet disparity economies and even small companies – in, say,
in the APAC region. In an article published by The Indonesia, Thailand, Myanmar, or elsewhere – to
Diplomat, she points out that “high income connect to the rest of the globe and compete. It can
countries [are] experiencing a higher growth rate of change the world and it will help shape the future.”49
broadband penetration relative to other countries.”
In short, more mature economies can devote more Despite the major opportunities it represents, cloud
resources to building and maintaining internet computing is not yet sufficiently implemented
infrastructure.46 across the APAC region to effectively mitigate the
region’s infrastructure and connectivity disparities.
In this sense, the growth of AI as a driving economic The Asia Cloud Computing Association (ACCA)
force hinges on the region’s ability to overcome the finds that while cloud infrastructure and expertise
inequalities that block its overall digital enablement. have become central components of many APAC
To realise the benefits of AI while navigating their economies’ wider national digital transformation
own specific economic challenges, APAC countries agendas, a cloud computing divide does exist – and
will need to promote an inclusive digital economy persist – between emerging and mature markets.50
and society, where none are left behind.
For the ACCA, the key differentiator will be
Cloud computing to bridge the divides countries’ adoption of ‘Cloud First’ policies –
frameworks designed to push organisations and
A recent Oxford Economics report notes that cloud institutions to consider cloud computing solutions as
can help bridge infrastructure and connectivity gaps a primary part of their IT architecture. Such policies
by putting sophisticated technologies within reach not only drive countries’ cloud adoption, they also
of smaller companies and emerging economies.47 contribute to their ability to harness the broader
technological advancements that cloud computing
Flexible and distributed cloud platforms give enables.
companies of all sizes access to cutting-edge data
tools, limitless storage, sophisticated software, and As Jarom Britton, ACCA Chair for Public Sector
the ability to scale up or down quickly according to Special Interest Group, notes in an online interview:
market factors – all without forcing them to invest “We always hear people talk about AI (artificial
in expensive servers or computers. This makes intelligence), blockchain, IoT (Internet of Things).
it easier for companies to experiment with new Well you don’t have AI if you don’t have cloud. You
services, expand their global reach, and build solid don’t have IoT if you don’t have cloud. You don’t
foundations for the next generation of technologies have e-commerce if you don’t have cloud… None of
that will drive business growth, including AI. those things can happen unless cloud computing is
facilitated.”51
Cloud platforms also shift the economics of IT in
favour of emerging economies. Cloud technology
standardises and pools IT resources and automates

Key Topic 2: Access to Data The public sector also benefits from cross-border
data flows. Organisations in the healthcare sector,
The free and secure flow of data across borders and for instance, increasingly rely on access to data sets
jurisdictions is crucial for the growth and expansion to improve research and provide faster, better health
of AI. This could include providing publicly available services. Data sharing has proven critical in saving
data, establishing data trusts, promoting data lives as well as preserving economic and social
portability, and facilitating cross-border data flows. It stability in the event of epidemics, such as the Ebola
is also important that legal and regulatory measures and the Zika outbreaks.53
designed today can evolve along with the AI
innovations of tomorrow. Across the board, the free flow of data accelerates
the spread of ideas, research, and technologies,
Data classification frameworks in which certain types leading to the emergence of new, innovative
of data may not be easy to design and implement, business models and opportunities that propel
but they do constitute a proven and implementable economies forward. But for a number of APAC
way forward. And, as Prime Minister Shinzo Abe governments, transnational data movements are
declared at the 2019 World Economic Forum, it is closely tied to rising privacy and security concerns;
in everyone’s best interest to put structured and namely, the perceived loss of control over where
nuanced cross-border data sharing at the top of data is held and where it can be transferred, as well
governments’ economic agendas.52 as the lack of visibility on who controls global data
flows and who is to be held liable and/or responsible
for any potential misuse.
There are also mounting concerns over the way AI

systems use the large amounts of available data to
make decisions autonomously – raising important
questions in terms of trust, responsibility, and
accountability that governments do not necessarily
feel equipped to address. To ensure they remain
in control, many governments enact measures
that restrict the cross-border movements of data,
inadvertently hampering digital businesses’ access
to global opportunities, and even negating the
benefits brought about by global connectivity.
Cross-border data flows enable digital innovations
Data localisation requirements are a barrier to AI
Cross-border data flows are an enabling force for
businesses around the world. From manufacturing Despite the significant benefits of cross-border
and services to agriculture and retail, all sectors data flows to companies, consumers, and national
increasingly rely on data – and on its global flow – economies, many countries in the APAC region have
to plug into global value chains and contribute to the measures in place that limit or restrain the flow of
global economy. data across borders.
SMEs, especially, benefit from the flow of data China has a number of regulatory measures
across borders; it helps reduce transaction restricting routine cross-border transfers of
costs and mitigate the constraints of geographic information. The Cybersecurity Law requires
distance, increasing overall efficiencies as well as many international data transfers to be subject
opportunities. Small businesses, like large ones, are to a security assessment by the relevant industry
able to reach global markets and leverage data- regulator. It also states that personal information
driven services to meet consumers’ fast-evolving and important data must be stored locally, especially
needs and expectations.

when companies operate in “critical information Data classification frameworks allow governments
infrastructure sectors”.54 to bring specific solutions to specific problems,
without adding unnecessary risks, costs, or burdens.
Indonesia, meanwhile, has a number of regulations Organising data by specific criteria – level of
that indirectly restrict the movement of data across sensitivity, degree of strategic value, or extent of
platforms and jurisdictions. Its data localisation exposure to threats – makes it possible to determine
requirements compel providers of certain electronic which data can flow without limit (business-enabling
services to use or establish data centres within data, for example), and which data may be clearly
Indonesia.55 Other regulations require certain identified as vital or sensitive such that it needs to
personal and financial data to be stored and be restricted in movement (data related to national
processed locally. Another regulation requires security, for instance).
providers of “over the top” (OTT) services –
essentially every service provided via the internet – By classifying information based on risk, an
to register with the government, identify permanent organisation can focus its security efforts more
local representatives, and open bank accounts in efficiently, mobilising the right resources to protect
Indonesia.56 the right information. This leads to better security, as
restrictions are put exactly how and where they are
In South Korea, the Personal Information Protection needed, without limiting businesses’ access to data
Act requires companies to obtain consent from in the process. Box 2 shows Australia’s approach to
“data subjects” (i.e., the individuals associated with data classification.
particular data sets) prior to exporting that data. The
Act also requires “data subjects” to be informed of
who receives their data, the recipient’s purpose for
having that information, the period that information
will be retained, and the specific personal
information to be provided.57 This can constitute a
substantial burden on companies trying to send data
across borders.
According to the Information Technology &

Innovation Foundation (ITIF), countries that
enact barriers to data flows make it harder and
more expensive for domestic companies to gain
exposure and to benefit from the ideas, research,
technologies, and best practices that accompany
data flows – including AI-enabled innovations.58
Data classification to balance security with data-

driven innovation
Governments do not have to choose between

enabling the flow of data across borders and
upholding privacy and security principles. Indeed,
several of the highest-ranking economies have
implemented – or are looking to implement –
regulations that structure cross-border data flows in
a more balanced, nuanced, and targeted manner.

Box 2. Australia’s Model Data Classification Framework
Data classification is a risk-based approach to the management of data, and is a key component of any
comprehensive security policy. It involves classifying data into categories based on sensitivity levels and
risk profiles, and then aligning security controls needed for each level to manage risks appropriately.
Typically, classification of information is based on the potential impact to the national interest,
organisations, or individuals, that would arise if the information’s confidentiality, integrity, or availability
are compromised.59
The fewer the levels of classification, and the clearer the distinctions between levels, the more effective
the classification system will be in ensuring the most sensitive information is properly secured.
Governments often employ a three-tiered classification scheme, with the majority of public sector data
classified in the two lowest tiers.
The Australian Government uses categorisation model to assess the sensitivity or security classification
of information by considering the potential impact that could arise from compromise of the information’s
confidentiality, integrity, and availability.
The Australian Government uses three security classifications:60
Top secret: Most valuable, important and sensitive information. Compromise of information would be
expected to cause exceptionally grave damage to the national interest, organisations, or individuals.
Secret: Very valuable, important and sensitive information. Compromise of information would be
expected to cause serious damage to national interest, organisations, or individuals.
Protected: Valuable, important and sensitive information. Compromise of information would be expected
to cause damage to the national interest, organisations, or individuals.
Other official information that does not warrant a security classification, is classified according to its
sensitivity: 61
Official: Sensitive: Sensitive information is that which would result in limited damage to an individual,
organisation or government if compromised.
Official: Other information from routine business operations and services, the compromise of which
would have a low impact.
Unofficial: Information which does not form part of official duty.
The majority of information that is created or processed by the Australian Government is considered
‘official’ information.
Source: TRPC Research

Key Topic 3: Skills and Human Capital gap. Yet few APAC countries have begun to address
the impact of automation through educational
In the digital economy, workers and businesses policy. Intelligent automation is expected to boost
looking to use and benefit from AI systems need the importance of both education related to STEM
computer science skills and knowledge. Equipping (science, technology, engineering, and mathematics)
populations with AI skills is key to solving two inter- and of so-called soft skills, which allow workers to
connected barriers to the potential benefits of AI: trade on their uniquely human capabilities.
the (real or perceived) risk of widespread job losses
due to AI, and the rising talent gap that businesses However, in all but the highest-scoring countries, little
face as their processes increasingly rely on AI-based has been done to prepare future workers through
systems. school curricula or, just as importantly, teacher training.
At the same time, some experts warn that a focus on
Preparing students for the workplace of the future soft skills would be a distraction in countries where
even basic education remains a challenge.65
According to recent estimates, about 15% of
work activities are likely to be displaced by 2030. In Singapore, AI skills are considered one of the four
The proportion varies widely across countries; core abilities that will help the country strengthen its
automation may negatively impact mature markets digital readiness,66 which is why the government has
– reflecting higher wage rates and thus stronger introduced experimental coding classes for all upper
economic incentives to automate – while it may primary pupils. Conducted by the Education Ministry
create unique leapfrogging opportunities for and the IMDA, the 10-hour enrichment programme
emerging markets.62 aims to develop an appreciation of core computational
thinking and coding concepts in pupils, and will be
A report on the advantages of automation finds rolled out to all primary schools by 2020.67 Singapore
that, if harnessed properly, mature economies is also experimenting with funding “individual learning
such as Australia stand to boost their economy by accounts”, which adults can use to support training
up to USD1.5 trillion (AUD2.2 trillion) by 2030.63 At courses throughout their lives.
the same time, however, economies that are less
equipped to transition into an AI-driven society face In South Korea, the MSIT announced a dynamic talent
a number of difficulties and drawbacks, including the development system to nurture 10,000 AI-ready
possibility of massive job displacements and the risk talents. A KRW180.6 billion (USD149 million) investment
of heightened socio-economic inequalities. into an Innovative Academy will train technical
capabilities of 2,500 students, provide overseas
In 2015, computer science professor and artificial postgraduate training for 2,250 students, set up AI
intelligence pioneer Stuart Russell wrote an open departments in graduate schools to train 860 talents,
letter calling for researchers to maximise AI’s social and provide industry-specific training to 7,000 talents
impact as it is made more capable and powerful: for ‘quick entry’ into the economy.68
“The potential benefits [of AI research] are huge,
since everything that civilization has to offer is a In Malaysia, AI, computer programming, and robotics
product of human intelligence; we cannot predict will be introduced as a compulsory subject to Year
what we might achieve when this intelligence Four students in 2020.69 Coding was introduced in
is magnified by the tools AI may provide, but secondary schools in 2017, via Basics in Computer
the eradication of disease and poverty are not Science (Asas Sains Komputer) and Computer Science
unfathomable. Because of the great potential of AI, (Sains Komputer). These subjects expose students to
it is important to research how to reap its benefits simple coding methods using Microsoft Visual Basic,
while avoiding potential pitfalls.”64 JAVA, HTML, Javascript, Microsoft Access, MySQL,
XAMPP, and Notepad. The Ministry of Education is also
Curriculum reform and lifelong learning are key working with MDEC to train lecturers at the Institutes
facets of education policy that can effectively of Teacher Education (IPG), as well as to provide
address both worker displacement and talent coding skills to about 500 teachers.70

Helping workers navigate new labour dynamics Washington state, while Alibaba is planning on opening
seven labs worldwide aimed at beating similar research
The challenge for government and industry, with facilities built by IBM, Microsoft, and Intel.76
the help of educational institutions, is to ensure that
workers are provided ample opportunity to gain Another way is to reinforce training initiatives aimed
the skills needed to operate effectively in the future at equipping existing technology workers with more
workplace and take advantage of opportunities advanced AI skills and knowledge. In Singapore,
brought about by automation. Labour market government agency AI Singapore launched a
policies to enable greater worker mobility and three-year programme, comprising workshops and
flexibility – for example, through relaxed licencing apprenticeship schemes, to equip 12,200 Singaporean
requirements for certain occupations or more professionals and engineers with marketable AI
portable workplace benefits – can help countries knowledge.77
meet this challenge.
Private companies are also creating re-skilling and
According to the EIU Automation Readiness Index, re-training opportunities. In Japan, multinational
the countries where such policies and programmes electronics manufacturer Daikin launched an internal
are closest to being in place are the same that are programme aimed at turning 1,000 employees with no
the most supportive of AI and robotics innovation, background in AI into skilled AI professionals. Following
and are beginning to address the associated specialised courses at Osaka University for six months,
educational challenges. In the APAC region, the selected employees are then assigned special
Singapore, South Korea, and Japan all rank highly in data-based roles within the company to hone their
this regard, demonstrating labour markets that are newly acquired skills.78
both strong enough to protect workers from severe
disruptions and flexible enough to prepare them for Universities are also playing a role in ensuring
the inevitable changes ahead.71 workers across the APAC region are not left behind.
According to Dr Norisma Idris, Associate Professor
Upskilling and retraining to mitigate the AI talent at the Department of Artificial Intelligence (University
gap of Malaya), the Malaysian government is “initiating
industry partnerships with universities, as well
Fears of widespread job losses due to AI are only as creating collaboration opportunities between
one side of the coin. Businesses where robotics researchers and government agencies. It is also
and automation are being actively deployed report launching and promoting new AI and Data Science
major gaps in AI-capable talent, making it difficult to programmes, as well as introducing AI-related courses
benefit from the transformative benefits of AI.72 This within other programmes. All this is to increase
is also the case in the APAC region, where only 1 in the number of students registering in AI-related
every 1,000 candidates is qualified to use, develop, programmes.”79
or run an AI system.73
In China, more than 70 Chinese universities and
According to a study conducted by Chinese tech colleges have introduced AI-related majors, and
giant Tencent, there are just over 300,000 AI 283 universities are licensed to offer data science
researchers and practitioners worldwide, when programmes.80 In India too, AI-related courses and
several millions are needed to adequately meet programmes are in high demand – so much so that
market demand.74 In China alone, demand for AI some states are finding it difficult to keep up with the
professionals may surge to 5 million in a few years, rising demand.81
with researchers in machine learning, smart chips,
and algorithms being the most sought-after.75 Ralph Haupter, President of Microsoft Asia, notes in
One way Chinese companies are addressing the a blog post, “In the AI era, everyone – at all levels, in
talent gap is to launch overseas labs to attract all roles, and especially from all cultures and walks of
foreign researchers and Chinese nationals living like – needs to be equipped with tools and skills to be
abroad. Tencent opened its own AI lab in Seattle, future-ready.”82

Key Topic 4: Trust and Partnerships tailored products and services.85 CCTV systems in
China,86 Malaysia,87 and Singapore88 are increasingly
Data-driven technologies such as AI need to be equipped with AI-based voice, facial, and even gait
developed in a safe, ethical, and sustainable manner recognition capabilities.89
for citizens and consumers to understand and trust
them. This is especially true for a growing number While all these systems are primarily designed to
of digital devices and services – including data- improve citizens’ and consumers’ lives, they do
intensive AI systems – that rely on personal data create massively detailed datasets that could be
to function. This means developing frameworks used for criminal activities if they were to fall in
that ensure AI systems are fair, transparent, and the wrong hands. The problem is that most trust
inclusive, and rooting these frameworks in local, and privacy regulations – where they exist – were
regional, and international collaboration schemes enacted before the use of the Internet of Things
that enable best practices, standards, and principles (IoT), big data, and AI were a consideration, and
to be shared globally. when consent provisions were defined more simply
than they are today. That is why it is preferable for
AI systems to clearly demonstrate that they comply
with data protection and privacy laws that regulate
how the data of individuals and organisations is
collected, stored, used, and disseminated.
This is an emerging topic all around the region, and

one that is yet to come to concrete conclusions.
APAC economies are in varying stages of
development when it comes to their respective data
protection and privacy regimes. Malaysia, Singapore,
Trust as a defining concept of data-driven societies and the Philippines have clear and thorough data
protection laws in place, equipping both businesses
According to a recent Salesforce report, APAC and governments to face data-related challenges in
consumers do not necessarily understand the the digital age.
way AI technologies work, despite being acutely
aware of them. This translates into an observable In Singapore, the Personal Data Protection
reluctance to trust AI technologies, even knowing Commission (PDPC) released a discussion paper
full well that they can potentially improve different titled “Artificial Intelligence (AI) and Personal Data –
aspects of their daily lives.83 Fostering Responsible Development and Adoption
of AI”, which proposes an accountability-based
A key issue is that the more data-intensive digital framework on consumer protection issues related to
platforms and devices become – and the more that AI. The proposed governance framework put forth
data is entrusted to automated, algorithmic decision- by the PDPC seeks to strike a balance between
makers – the more difficult it becomes for citizens maximising the benefits of AI without compromising
to know whether their data is used for the purposes privacy and accountability.
and in the manner for which it was collected (i.e. that
their data is responsibly and ethically used). In Thailand, the National Legislative Assembly
approved the Personal Data Protection Act, B.E.
For example, Airbnb leverages TensorFlow, Google’s 2562 (2019) (“PDPA”).90 As the country’s first
open-source library that supports machine learning, consolidated law on data protection, it frames the
to identify and predict users’ accommodation collection, use, and disclosure of personal data,
preferences.84 Singapore’s OCBC bank developed drawing key concepts and principles from the
AI solutions that learn from customer data – EU General Data Protection Regulation (GDPR)
purchasing behaviour as well as interactions with – though it is yet unclear if the law has specific
relationship managers – to offer targeted and provisions for data used by AI systems.

Ethics as a driving principle of AI-based systems Analytics in Singapore’s Financial Sector”.97 The
government also convened an Advisory Council on
Beyond regulatory measures to instil trust, driving AI the Ethical Use of AI and Data, comprising experts
readiness and adoption also requires accountability. in AI and big data from local and international
For instance, when an AI application is in charge companies, as well as academia and consumer
of deciding who qualifies for a loan, it is important advocates.98
for the bank to be able to explain the algorithm’s
decision, as well as take full responsibility in case While not necessarily going to the same lengths
of errors or oversights.91 This entails designing AI as Singapore, most other governments with
systems that are fair, reliable, and safe, as well as AI strategies are also employing the language
secure, inclusive, transparent, and accountable. of ethics to foster the responsible use of AI.
Singapore’s lead in grappling with issues of
So far, there are no specific laws governing the explainable, transparent, and fair algorithms, as well
ethical use of AI in the region, but efforts are as in practically incorporating considerations for
underway to better understand and address legal, competition, privacy, and ethics into its policy and
moral, and ethical issues raised by the spread of AI regulatory frameworks, will be a useful resource for
systems and services across economic sectors.92 other governments as they formulate their own AI
strategies.
In Japan, the Ministry of Economy, Trade, and
Industry (METI) formulated the “Contract Guidance The difficulty lies in two different, but related, facets
on Utilisation of AI and Data” to help organisations of AI. First, the fact that the more data is created
navigate the complexities of drafting a contract that and fed into AI systems, the more difficult it is to
involves AI- and data-based systems.93 METI also guarantee that it will be strictly used in the manner
announced more in-depth guidelines this year to or for the purpose for which it was collected. In
address issues such as legal liability and user rights. this sense, moral principles and ethical guidelines
China, meanwhile, released the Beijing AI Principles, must be instilled into the AI systems themselves,
a code of ethics for AI that aims to guide AI as well as into the regulations that frame them.
scientists and engineers as they research, develop, Box 3 presents Singapore’s approach to making AI
use, and plan AI-based systems.94 both ethical and trustworthy. Second, the fact that
a full and unequivocal comprehension of human
In Australia, the Commonwealth Scientific and decision-making, including heuristics, biases, group
Industrial Research Organisation (CSIRO) launched dynamics, creativity, and imagination must first be
a landmark discussion paper on a proposed Ethics attained before attempting to use them to improve
Framework.95 The paper highlights eight core computer forecasting models, predictive analytics,
principles to guide AI deployment in the Australian or robotic systems.
economy: generating net-benefits, doing no harm,
complying with regulatory and legal requirements,
appropriately upholding privacy, boosting fairness,
being transparent and easily explained, containing
provisions for contesting a decision made by a
machine, and including an accountability trail.
In Singapore, the Ministry of Transport set up a

Committee on Autonomous Road Transport, which
will look into regulating the use of driverless cars in
the near future.96 In 2018, the Monetary Authority of
Singapore (MAS) released the “Principles to Promote
Fairness, Ethics, Accountability and Transparency
(FEAT) in the Use of Artificial Intelligence and Data

Box 3. Singapore’s Model AI Governance Framework
The Personal Data Protection Commission (PDPC), Infocomm Media Development Authority
(IMDA), with the advice from the Advisory Council on the Ethical Use of AI and Data recently
published the Proposed Model AI Governance Framework (Model Framework).
The Model Framework translates ethical principles into practical measures that can be
implemented by organisations deploying AI solutions at scale. One of its aims is to promote AI
adoption while building consumer confidence and trust in providing their personal data for AI-
focused purposes.
The Model Framework is based on two high-level guiding principles that promote trust in AI and
understanding of the use of AI technologies:
1. Decisions made by or with the assistance of AI should be explainable, transparent, and fair
so that affected individuals will have trust and confidence in these decisions.
• Explainable: Automated algorithmic decisions and the data that drives such decisions must
be understandable by end-users and other stakeholders in non-technical terms.
• Transparent: AI developers, data scientists, application builders, and companies should
be accountable for the AI algorithms, systems, applications, and resultant decisions
respectively in order to build trust in the entire AI ecosystem.
• Fair: AI algorithms and models embedded in decision-making systems should incorporate
fairness at their core to avoid unintentional biases or discriminatory practices.
2. AI systems, robots, and decisions made using AI should be human-centric, i.e. put users
front-and-centre of all and any AI deployment.
• Decisions should strive to help individuals and avoid causing foreseeable harm.
• Tangible benefits to individuals should be identified and communicated to build consumer
understanding and confidence.
• AI systems and robots should be designed to avoid causing bodily harm or affecting the
safety of individuals.
Source: Personal Data Protection Commission (PDPC), www.pdpc.gov.sg/-/media/Files/PDPC/PDF-Files/Resource-for-Organisation/AI/A-

Proposed-Model-AI-Governance-Framework-January-2019.pdf

Key Topic 5: Ecosystem and well-funded national AI strategies with the clear
Entrepreneurship objective of maximising the potential benefits of
AI.104 From USD21.6 million in Australia (Australian
Building digital economies and digital societies is Technology and Science Growth Plan) and USD91.5
key to competitiveness in the digital age. Across million in Singapore (AI Singapore) to USD1.95 billion
the APAC region, digital economy initiatives in South Korea (Artificial Intelligence R&D Strategy),
encompass a wide range of AI-enabling efforts; from governments are clearly focusing their resources on
investments in AI start-ups to matching business ramping AI capabilities.
problem statements with AI research talent,
governments across the region are developing China stands out in this regard. Determined to
a broad base of AI expertise to strengthen and become a global AI powerhouse by 2030, the
consolidate homegrown AI ecosystems. Chinese government is said to have invested USD12
billion in AI in 2017 – and is likely to invest at least
Investing in innovation USD70 billion by 2020.105 These funds are used to
develop AI projects for the government, but also to
Across the board, private- and public-sector support the rapid growth of domestic AI companies.
investments are driving countries’ ability to research, Other initiatives include the announcement of a
develop, and apply AI technologies. IDC estimates USD2.1 billion-blueprint to build an AI industrial
that global spending on cognitive and AI systems park in the suburbs of Beijing,106 and the Ministry of
will reach USD77.6 billion in 2022, more than Education’s launch of a five-year AI talent training
three times the USD24 billion forecast for 2018 – a programme.107
compound annual growth rate (CAGR) of 37.3%.99
The same estimates see software as both the China’s race to the top via generous investments
largest and fastest growing technology category, seems to be paying off. According to the Allen
representing around 40% of all cognitive/AI Institute for Artificial Intelligence, China’s output of
spending with a five-year CAGR of 43.1%. Two areas influential AI research papers is close to overtaking
of focus for these investments are conversational that of the United States, the world’s current number
AI applications (e.g., personal assistants and one in AI research. Measuring not just the number
chatbots) and deep learning and machine learning of papers, but how often they are cited – a good
applications (employed in a wide range of use shorthand measure for influence in the wider scientific
cases).100 community – the Allen Institute finds that China is
poised to overtake the United States in the 1% of
Big tech and other venture capital firms are also most-cited papers by 2025.108
focusing their investments on AI. Pitchbook, an
investment analyst firm, estimates that in 2017, USD6 Building conducive environments
billion was invested across 643 venture capital
deals in AI.101 According to McKinsey estimates, In addition to investment, policies, laws, and
multinational corporations such as Alibaba, Amazon, regulations play an essential role in enabling the
Google, Baidu, and Facebook invested between many AI initiatives that start-ups, entrepreneurs,
USD20 billion and USD30 billion globally in AI in research laboratories, and academics prepare and
2016.102 From 2011 through February 2017, these launch every year.
companies were behind 29 of 55 major merger &
acquisition deals in the United States and nine of ten As mentioned earlier in this report, cybersecurity and
major deals in China – an investment strategy that privacy protection frameworks are key elements of
is helping these companies acquire pivotal talent, a conducive regulatory environment. But there are
technology, and data sets, potentially raising barriers many other items at play. Namely, all the measures
for slower-moving competitors.103 that allow experimental AI projects to become
marketable products and services, with real-life
Governments, meanwhile, are also doing their part. applications as well as sustainable business models.
Since 2017, at least 20 countries have released Speaking at an MIT conference on AI policy, Regina

Barzilay, MIT professor at the Computer Science consumers. One key feature is matching research
and Artificial Intelligence Laboratory (CSAIL), cites talent in research institutions with specific industry
the example of AI in healthcare. Her research problems.”114
uses machine learning to analyse mammogram
results for better early detection of breast cancer, Writing in Ethos, Ng Chee Khern, Permanent
but the limited access to data and to large Secretary for Smart Nation and Digital Government,
testing populations makes it challenging to refine echoes this sentiment: “We are witnessing a
automated detection tools.109 R. David Edelman, paradigm shift where [government] agencies are
Director of the Project on Technology, the Economy, beginning to see technology not just as an expense,
and National Security (TENS) at the MIT Internet but also as an investment in new strategic capability:
Policy Research Initiative (IPRI), notes during the that it is about how technology can help agencies
same event that this is an area that public policy reach top-line growth (faster time-to-market for
could and should address, as governments are best services, improved service delivery) in achieving
positioned to bridge discussions between citizens, mission objectives.”115
policy-makers, and artificial intelligence technical
communities.110 In this context, several coordinated programmes
are in place to ensure AI effectively addresses as
This propensity to build networks and promote many socio-economic challenges as possible. AI
collaboration is a natural strength for Singapore, one Singapore, the national programme for developing
of APAC’s most prominent AI-enabled economies. artificial intelligence capabilities, has the Grand
A historical hub for regional and global flows of Challenge initiative – a five-year, USD109 million
capital, people, technology, and ideas, it has built investment fund for developing AI solutions to solve
strong organisational and institutional capabilities, the nation’s most pressing issues in healthcare,
allowing it to craft a conducive environment finance, and urban mobility. There is also the 100
for AI-focused innovation to grow and mature. Experiments programme, a business-focused
Singapore’s thriving AI sector is thus the product of initiative that helps companies use AI to solve
a range of converging factors: a prime location at day-to-day issues, including operational efficiency,
the crossroads of Asia’s biggest markets, low taxes product design, and customer engagement.116
on businesses, a well-developed IT infrastructure,
strong investment opportunities, and a robust Building a vibrant core of AI developers and
regulatory regime. companies is another goal of IMDA’s strategy. Four
key initiatives can be highlighted: a nine-month AI
Through the Infocomm Media Development Apprenticeship Programme for recent graduates;
Authority (IMDA), Singapore has coordinated a three-month AI for Industry programme for IT
various AI efforts undertaken by ministries, professionals and software developers; AI for
agencies, academia, and a growing collection Students, an online learning programme for schools;
of homegrown and multinational AI firms. This and AI for Everyone, a short course that will give
includes sector-specific initiatives launched by the 10,000 non-technical Singaporean workers basic
National Research Foundation (industry-driven AI familiarity with how AI can apply to their jobs and
applications),111 the Economic Development Board lives.117
(productivity-focused programmes),112 and the
Monetary Authority of Singapore (fair and ethical AI Building an AI ecosystem in such a comprehensive,
in the financial sector).113 consistent, and coordinated manner may not yet
be within reach of all APAC governments. The
Interviewed in a MIT Technology Review report, example set by Singapore may serve as a guide for
Zee Kin Yeong, Assistant Chief Executive, Data governments looking to holistically and sustainably
Innovation and Protection at IMDA, says that strengthen their AI capabilities.
the overarching goal for Singapore “is to see
how Singapore-based AI research can filter into
products and into the hands of companies and


IV.
The Policy
Response to AI

Overview of AI Strategies
While AI is an emerging, fast-evolving technology, some governments are

making major strides in the way they are designing frameworks around it.
From overarching national AI strategies to sector- to emphasise the importance of working together
specific guidelines for AI implementation, APAC to ensure AI is developed according to their shared
governments are embracing and leveraging AI at values.119
their own pace and in their own manner.
Although some EU countries have their own national
The AI frameworks described in this section cover AI initiatives, the Declaration reinforces the need to
a selected range of AI-specific initiatives that are work together to enhance research and deployment
being developed and implemented globally, as well while dealing collectively with social, economic,
as across the APAC region.118 ethical, and legal questions.
At the same time, the European Commission

Global AI Policy Strategies (EC) adopted the Communication on Artificial
Intelligence, a 20-page policy document that lays
European Commission out its approach to AI.120 Through this, the EC aims
to:
The European approach to AI emphasises three
key areas: collaboration across European Union 1. Increase the EU’s technological and industrial
(EU) members, the promotion of shared values, and capacity and AI uptake by the public and private
the development of ethical guidelines and best sectors;
practices. 2. Prepare Europeans for the socioeconomic
changes brought about by AI; and
In April 2018, 25 European countries signed a 3. Ensure that an appropriate ethical and legal
Declaration of Cooperation on Artificial Intelligence framework is in place.

Key initiatives include: 6. Societal and environmental well-being: AI
systems must consider the broader society,
Investment: A commitment to increase the EU’s other sentient beings, and the environment as
investment in AI from EUR500 million in 2017 to stakeholders throughout their life cycle.
EUR1.5 billion by the end of 2020; 7. Accountability: AI systems must be designed
in a way that ensures responsibility and
Cooperation: The creation of the European AI accountability for the systems and their
Alliance and the High-Level Group on Artificial outcomes, both before and after their
Intelligence; development, deployment, and use.
Principles: The development of AI ethics guidelines

for member states to address issues such as In June 2019, during the first AI Alliance Assembly,
fairness, safety, and transparency; the High-Level Expert Group on Artificial
Intelligence (AI HLEG) presented its Policy and
Investment Recommendations for Trustworthy
In April 2019, the High-Level Group on Artificial
Artificial Intelligence.122
Intelligence published the final Ethics Guidelines for
Trustworthy Artificial Intelligence, which lists seven
The report lists 11 key recommendations aimed
key requirements that AI systems should meet in
at helping EU members go beyond national AI
order to be considered trustworthy by European
strategies and work towards a concerted, human-
standards:121
centric approach to AI:
1. Human agency and oversight: AI systems

• Empower and protect humans and society
should support human autonomy and decision-
• Take up a tailored approach to the AI landscape
making, as prescribed by the principle of respect
• Secure a Single European Market for
for human autonomy.
Trustworthy AI
2. Technical robustness and safety: AI systems
• Enable AI ecosystems through Sectoral Multi-
must be developed with a preventative
Stakeholder Alliances
approach to risks and in a manner such that they
• Foster the European data economy
reliably behave as intended while preventing
• Exploit the multi-faceted role of the public sector
unintentional, unexpected, and unacceptable
• Strengthen and unite Europe’s research
harm.
capabilities
3. Privacy and data governance: AI systems
• Nurture education to the Fourth Power
necessitate adequate data governance that
• Adopt a risk-based governance approach to
covers the quality and integrity of the data used,
AI and an ensure an appropriate regulatory
its relevance in light of the domain in which
framework
the AI systems will be deployed, its access
• Stimulate an open and lucrative investment
protocols, and the capability to process data in a
environment
manner that protects privacy.
• Embrace a holistic way of working, combing a
4. Transparency: AI systems must ensure that all
10-year vision with a rolling action plan
elements – the system itself, the data it uses,
and the business models it works towards
– work in accordance with the principles of
transparency, traceability, and explainability.
5. Diversity, non-discrimination, and fairness: AI
systems must enable inclusion and diversity
throughout their life cycles, ensuring equal
access through inclusive design processes as
well as equal treatment of all stakeholders.

G20 OECD
In June 2019, the G20 – of which Australia, In May 2019, the OECD – of which Australia, Japan,
China, India, Indonesia, Japan, and South Korea New Zealand, and South Korea are members –
are members – adopted a set of non-binding AI adopted the OECD Principles on AI to help member
Principles to ensure AI is developed in a human- economies formulate consistent and concerted
centric manner.123 public policies and strategies on AI.125
Overall, the G20 guidelines call for users and The responsible stewardship of trustworthy AI is
developers to ensure AI is fair and accountable, guided through five main complementary, values-
with transparent decision-making processes and based principles:
that respects the rule of law and values including
privacy, equality, diversity, and internationally • AI should benefit people and the planet
recognised labour rights. The principles also urge by driving inclusive growth, sustainable
governments to ensure no citizen is left behind as development, and well-being.
AI advances, providing a fair transition for workers • AI systems should be designed in a way that
through training programmes and access to new job respects the rule of law, human rights, democratic
opportunities. values, and diversity, and they should include
appropriate safeguards – for example, enabling
Based on the OECD’s own OECD Principles on AI, human intervention where necessary – to ensure
the G20 principles underline the need for: a fair and just society.
• There should be transparency and responsible
• Inclusive growth; disclosure around AI systems to ensure that
• Sustainable development and well-being; people understand AI-based outcomes and can
• Human-centred values and fairness; challenge them.
• Transparency and explainability; • AI systems must function in a robust, secure,
• Robustness; and safe way throughout their life cycles and
• Security and safety; and potential risks should be continually assessed
• Accountability. and managed.
• Organisations and individuals developing,
A number of guidelines are also provided for G20 deploying, or operating AI systems should be
policy-makers to facilitate international cooperation, held accountable for their proper functioning in
capacity building, and research and investment in line with the above principles.
the field of AI.
Discussions are also ongoing on the launch of an

International Panel on Artificial Intelligence (IPAI).
Primarily championed by Canada and France,
the IPAI will bring together global AI experts to
facilitate and foster international collaboration and
coordination on AI policy developments.124

In line with these value-based principles, the OECD In April 2016, the Strategic Council for AI Technology
also provides five non-binding recommendations to was established to develop a roadmap for the
governments: industrialisation of artificial intelligence.128 The
11-member council included representatives from
academia, industry, and government, as well as the
• Facilitate public and private investment in President of Japan’s Society for the Promotion of
research & development to spur innovation in Science, the President of the University of Tokyo,
trustworthy AI. and the Chairman of Toyota.
• Foster accessible AI ecosystems with digital
infrastructure and technologies and mechanisms In March 2017, the Artificial Intelligence Technology
to share data and knowledge. Strategy was released along with an Industrialisation
• Ensure a policy environment that will open the Roadmap, which organises the development of AI
way to deployment of trustworthy AI systems. into three phases:129
• Empower people with the skills for AI and support
workers for a fair transition. 1. The utilisation and application of data-driven AI
• Co-operate across borders and sectors to developed in various domains;
progress on responsible stewardship of 2. The public use of AI and data developed across
trustworthy AI.126 various domains; and
3. The creation of ecosystems built by connecting
The OECD is also planning on launching an AI multiplying domains
Policy Observatory to facilitate dialogue and share
best practices on AI policies. A multi-disciplinary The strategy applies this framework to three
organisation, the AI Policy Observatory will help, priority areas of Japan’s Society 5.0 initiative130 –
among other missions, develop metrics to measure productivity, health, and mobility – and outlines key
AI research, development, and deployment, as well policies related to R&D, talent, public data, and start-
as gather evidence to assess members’ progress ups.
as they implementation the AI principles and
recommendations.127 In July 2017, Japan published the Draft AI R&D
Guidelines for International Discussions in
preparation for the Conference toward AI Network
Society.131 The document is not intended to be
a regulation, but a set of non-binding AI R&D
principles and guidelines promoting the benefits
and reducing the risks of AI.
The document includes the following Basic

Philosophies:
• Human-centred society
• Share guidelines as non-binding soft law with
Japan stakeholders internationally
• Ensure balance of benefits and risks
Japan was among the first countries to develop • Avoid hindering technologies or imposing
a national AI strategy. Largely focused on making excessive burdens on developers
AI useful to both society and economy, Japan’s • Review the Guidelines constantly and renew
approach aims to strengthen AI R&D capabilities, them as necessary
develop AI systems with industrial applications,
and prepare workers for the transformations of the
labour market.

Japan’s self-identified weakness in AI is its lack 4. Protecting American values including civil
of AI-enabled talent. The 2017 White Paper on liberties and privacy and fostering public trust
International Economy and Trade noted that there and confidence in AI technologies; and
is a talent shortage of approximately 50,000 in 5. Protecting US technological advantage in AI,
the technology field, including AI.132 In June 2019, while promoting an international environment
Prime Minister Shinzo Abe unveiled a plan to train that supports innovation.
250,000 people in AI skills annually by 2025, but
the plan has been criticised as unrealistic due to a The Executive Order calls on the National Science
shortage of AI-capable teachers.133 and Technology Council (NSTC) Select Committee
on Artificial Intelligence to coordinate the American
AI Initiative.
All executive departments and agencies that are

developing or deploying AI, providing educational
grants, or regulating or guiding AI are required to
adhere to six strategic objectives:
• Promoting sustained investment in AI R&D;

• Enhancing access to Federal data, models, and
computing resources;
• Reducing barriers to the use of AI technologies;
• Ensuring that technical standards minimise
United States
vulnerability to attacks from malicious actors;
• Training American AI researchers; and
In February 2019, President Donald Trump issued
• Implementing an action plan to protect US
the Executive Order on Maintaining American
economic and national security interests.
Leadership in Artificial Intelligence, effectively
launching the American AI Initiative.134

The Executive Order emphasises the central role
that the Federal Government plays in facilitating
AI R&D, training workers for a changing world,
promoting trust in AI, and using AI to protect national
interests, security, and values.135
And while the Executive Order emphasises

American leadership in AI, it is stressed that this
requires enhancing collaboration with foreign
partners and allies.
The American AI Initiative is guided by five

principles:
1. Driving technological breakthroughs;

2. Driving the development of appropriate
technical standards;
3. Training workers with the skills to develop and
apply AI technologies;

APAC AI Policy Strategies industries with AI and foster a domestic AI
market. These initiatives can be part of wider
The seven APAC economies covered in this report national technology plans, but they can also
are approaching AI in three distinct – but not address the specific needs and priorities of a
incompatible – ways: given sector.
• National / overarching AI strategies: Clear, • Foundational Guidelines, Principles, and

coherent, and comprehensive policies are Standards on AI: Another approach is to focus
essential to support economically strategic less on the pace and scope of AI growth and
areas. This is also true for AI. A complex, multi- more on fostering a conducive environment
faceted technology, accelerating its adoption at for a dynamic and sustainable AI ecosystem to
the national level requires whole-of-government emerge. Beyond the mobilisation of resources,
plans aimed at guiding and accelerating this entails ensuring that infrastructural,
concerted efforts, coordinated initiatives, and institutional, and organisational capabilities are
sustained collaboration across both existing in place, as well as defining the values, best
institutions and newly formed government practices, and technical standards that will allow
bodies. AI to prosper sustainably.
• Sector-Specific / Industry-Driven AI Strategies: It is worth noting that the categorisation within

The absence of a national AI policy does not these three major groupings is neither clear-cut
mean that nothing is being done to advance AI. nor definitive. Indeed, many of the APAC countries
In fact, most governments across the region have examined have initiatives and plans that run parallel
programmes to both upgrade existing industries to one another or that complement each other by
with AI and foster a domestic AI market. These addressing different specific needs simultaneously.
include programmes to both upgrade existing
National / Overarching AI Strategies
Country National AI Strategy
Australia None
China Next Generation Artificial Intelligence Development Plan
Indonesia None
Malaysia National AI Framework (forthcoming, announced in 2018)
Singapore AI Singapore, National AI Strategy
Mid-to Long-term Master Plan in Preparation for the Intelligent Information

South Korea
Society (forthcoming, announced in 2016)
Thailand None

Only China and Singapore have broad, overarching Malaysia and South Korea have national AI policies
national policies specifically designed to drive AI in the making, though in both cases there is little
adoption and to promote its deployment across information on how existing capabilities will be
public and private organisations. In both cases, the consolidated under an actionable, nationwide AI
government is launching multi-faceted initiatives agenda. Based on their current approaches to AI,
aimed at raising AI investment, supporting AI-based cooperation and coordination are expected to play
start-ups, enabling AI-capable manpower, and a major role in their plans, albeit in very different
improving security frameworks. manners.
The major difference between the two strategies Malaysia has so far prioritised cooperation with
is that Singapore is working closely with AI the private sector, multiplying the number of joint
researchers and academics to ensure these experimental projects that draw on the expertise
activities are conducted in an ethical, transparent, – and funds – of global AI players.138 South Korea,
and accountable manner,136 while China is adopting meanwhile, emphasises cooperation with academia
a more permissive, business-first approach with and research organisations, highlighting the
little to no oversight on the way AI ventures – data importance of skills, training, and knowledge in the
mining, location tracking, facial recognition – are maturation of its AI ecosystem.139
developed or deployed.137
Sector-Specific / Industry-Driven AI Strategies
Country Sector-Specific AI Initiatives
Australian Technology and Science Growth Plan, Cooperative Research Centres

Australia
(CRC) programme,140 AI PhD scholarships
China Artificial Intelligence Industry Innovation Alliance (CAIIIA), Three-Year Action

China Plan for Promoting Development of a New Generation Artificial Intelligence Industry
(2018–2020), AI Innovation Action Plan for Colleges and Universities
Indonesia Making Indonesia 4.0, Tokopedia AI Research Centre, NVIDIA AI R&D Centre
Industry Digitalisation Transformation Fund (2019 budget), Malaysia City Brain

Malaysia
project, Data Free Trade Zone (DFTZ), National Policy on Industry 4.0 (Industry4WRD)
Digital Economy Framework for Action, Autonomous Vehicle Rules, Principles to

Promote Fairness, Ethics, Accountability and Transparency (FEAT) in the Use of
Singapore
Artificial Intelligence and Data Analytics in Singapore’s Financial Sector, Veritas
framework
South Artificial Intelligence Information Industry Development Strategy, I-Korea 4.0 (2018-
Korea 2022), AI R&D Strategy, Data and AI Economy Revitalisation Plan (2019-2023)
Thailand Thailand 4.0, Digital Park Thailand, Thai People Map and Analytic Platform (TPMAP),

The approaches of Indonesia and Thailand stand
out in that they are both focused on industry-driven
AI. In both cases, AI policies are being launched to
drive the AI readiness and adoption of key economic
sectors (factories, farms, services, e-commerce),
with special focus on preparing small and medium
enterprises (SMEs) to capitalise on the opportunities
brought about by the “4.0 revolution”.141
A key challenge for both markets is ensuring their

institutional complexities do not get in the way of
regulatory clarity and consistency.142 The fact that
AI cuts across many strategic verticals may make it
easier to coordinate laws, policies, plans, agencies,
and bodies at all levels of government, avoiding
overlap and providing clarity to investors, business
leaders, and government stakeholders alike.
Foundational Guidelines, Principles, and Standards on AI
Country Foundational AI Plans and Programmes
AI Ethics Framework, AI Standardisation Roadmap, AI Technology Roadmap,

Australia
Humanising Machine Intelligence (HMI) project
China Beijing AI Principles
Indonesia Indonesia AI Forum on Data Privacy and Protection, Data Protection Bill (draft)143
Ministerial Value Innovative Centres (VICs, proposed),144 National Digital Identity

Malaysia
(forthcoming)
Advisory Council on the Ethical Use of AI and Data, five-year Research
Singapore Programme on the Governance of AI and Data Use, Proposed Model AI Governance
Framework145
South Korea KAIST Taming Artificial Intelligence: Engineering, Ethics, and Policy forum146
Thailand Personal Data Protection Act, Thailand Cybersecurity Act, Ethical Guidelines
Thailand
for AI (draft)
Australia has several federal and local plans that The private sector, meanwhile, strongly
prioritise both theoretical and applied research acknowledges the importance of AI in elevating
to enable the advancement of fair, inclusive, and the country’s economic competitiveness. A Deloitte
socially responsible technologies. Australia’s Chief survey finds that 50% of Australian executives see AI
Scientist, Dr Alan Finkel, has for example proposed currently being used to "catch up" with competition
an AI certification scheme to ensure AI companies rather than to establish a distinct advantage,148 a
meet relevant ethical standards and independent situation that has pushed industry leaders to call for
auditing requirements.147 greater AI spending and investment.149

Australia National AI Strategy No
Department of Industry, Innovation, and Science;
National Agency Commonwealth Scientific and Industrial
Research Organisation (CSIRO)
Data61, Credit Clear, Oovvuu, Hyper Anna,

Key AI Organisations
iCetana
Agriculture, Human/Social Services, Healthcare,

Key Industry/Sector
Defence, Financial Services, Mining, Oil and Gas
Machine Learning, Advanced/Predictive

Key AI Technology Analytics, Image/Pattern Recognition, Moral
Machine Intelligence (MMI)
Rank in EIU’s
Automation Readiness 10 / 25
Index150
Rank in Oxford Insights’
Government AI 11 / 194
Readiness Index151
Key Policy Key Agency Key Sector Key Objective
Key projects include:

• Upgrade supercomputers,
Australian • Support the Medical Industry Growth Plan,
Technology Department • Increase women’s participation in STEM, The government will allocate
and Science of Industry, • Establish a National Space Agency, AUD2.4 billion (USD1.63 billion)
Growth Innovation, • Upgrade satellite imagery and GPS to technology and science over
Program and Science systems, the next 12 years.
(2018) • Modernise patents management system,
• Strengthen AI and machine learning
capabilities.
Key areas of focus:

A roadmap for the government,
• People: Develop digital skills
the private sector, and the
• Services: Government to deliver better
community to develop world-
Digital Department digital services
leading digital business,
Economy of Industry, • Digital assets: Build infrastructure and
address the skills gap, and
Strategy Innovation, provide secure access to high quality data
equip workforces with the
(2017) and Science • Enabling environment: Maintain
necessary digital skills to
cybersecurity and review regulatory
support growth and job
systems to ensure they are fit for purpose
creation.
and technology neutral
National Four pillars: Promote and foster science

Innovation Innovation • Culture and capital; and innovation to generate
and Science and Science • Collaboration; insights and ideas, create jobs,
Agenda Australia • Talent and skills; and and drive long term economic
(2015) • Government as an exemplar. growth.

“Machine learning requires a lot
of curated data, and bigger, more
establiashed companies typically
do not have their data warehouses
set up in such a way that it is easy
to amalgamate all of the necessary
data… In contrast, start-ups develop
their business models with data
acquisition and curation in mind...”
Dr Zygmunt Szpak, Senior Research Associate, Australian

Institute for Machine Learning (AIML)
Country Overview and establishes the National Data Advisory Council,

a government agency overseeing the ethical use of
Australia does not yet have a broad, nationwide data.155
policy on AI, but the Department for Industry is
reportedly developing a national AI policy.152 Meanwhile, the 2018 Australian Technology and
Science Growth Program specifically earmarks
For now, however, government-led AI initiatives AUD29.9 million (USD21.6 million) over four years
are primarily public consultations that facilitate to strengthen the country’s AI capabilities. Of this
and open discussions on AI, aiming to enable the amount, AUD25 million will be used to support
development of official guidelines, principles, and industry-led Cooperative Research Centre (CRC)
standards to help public and private organisations Projects that focus on machine learning.156 The rest
brace for AI-driven disruption. Australia does, of the budget will be devoted to post-graduate
however, have several all-encompassing national scholarships, the development of a Technology
technology plans that recognise the economic Roadmap, a national AI Ethics Framework, and an AI
importance of AI and other advanced technologies, Standards Framework.157
but none of these constitute an AI-specific
framework. In terms of adoption, Dr Zygmunt Szpak, Senior
Research Associate at the Australian Institute
The National Innovation and Science Agenda for Machine Learning AIML), sees “Most of the
(NISA), for instance, is an all-encompassing national private sector activity is happening in relatively
technology plan that puts emerging technologies new companies or start-ups. Machine learning
such as cloud and AI at the heart of its agenda. requires a lot of curated data, and bigger, more
Among its main objectives are building world- established companies typically do not have their
class national research infrastructure, fostering data warehouses set up in such a way that it is
international research collaboration, incentivising easy to amalgamate all of the necessary data. It is
investment in innovative start-ups, and attracting quite a bureaucratic undertaking for them to get
and nurturing talent.153 In a similar vein, the Digital on top of what data they have, and understand its
Economy Strategy focuses on how the country quality. In contrast, start-ups develop their business
can leverage the economic benefits brought about models with data acquisition and curation in mind
by the digital economy.154 The strategy does not and hence, it is easier for them to explore machine-
specifically target AI, but it includes measures that learning solutions.”158
strengthen the country’s data analytics capabilities,

Dr Zygmunt Szpak adds that government adoption standards and ethics framework in place to support
is just as strong in Australia. “Many government the responsible development of AI, as well as
departments are exploring, adopting, or actively ensuring that the benefits of AI technologies
using AI technologies. For example, machine outweigh the impact on society. To this extent, the
learning is being used to audit tax returns and Australian government launched various public
predict congestion on traffic networks. Government consultations to facilitate and open discussions
is also promoting the adoption of machine learning on AI. Though unable to compete with the likes of
technologies by funding free workshops for various China and the United States when it comes to sheer
industry bodies. There seems to be a genuine AI investment, Australia can make a significant
interest in machine-learning technology form diverse contribution to the global advancement of AI. It
public-sector organisations. The Department of can, for instance, continue to develop the ethical
Defence is also investing heavily in adopting the principles, standards, and guidelines that will
latest AI techniques.”159 help worldwide AI initiatives keep AI ethical and
transparent.
Despite these initiatives, Deloitte’s State of AI in the
Enterprise survey finds that Australian businesses Cross-sectoral research collaboration between
are mainly using AI to “catch up” to competitors government, academia, and industry
rather than to “leapfrog ahead”.160 According to the
survey, 41% of Australian executives reported that In spite of not having a dedicated national AI
their company either completely lacks an AI strategy strategy there have been a number of organisations
or has only disparate departmental strategies, and individual experts engaging in AI research
compared to 30% of executives globally. Likewise, that stand to provide insights and deliver inputs
a recent BCG survey found that Australians are into a national framework (if and when the federal
generally supportive of AI, but their level of support government decides to develop one). Major
varies significantly by use case.161 For example, universities, as well as some corporations, a
there is overwhelming support for AI in areas such handful of start-ups, and state agencies are funding
as transport optimisation, predictive maintenance research and supporting AI pilots, in particular the
of public infrastructure, and customer services. government-supported CSIRO and Data61.
Conversely, 56% of people do not approve of using
AI to determine innocence or guilt in a criminal trial, Data61 and Monash University have, for example,
and 53% disagree with its use for making parole developed a ‘data airlock’ solution which uses AI
decisions. to analyse potentially harmful data and disturbing
images while protecting investigators from exposure
In this context, the lack of an overarching national to confronting images so they can carry out analysis
AI strategy may hinder the growth of Australia’s AI in a safe and secure manner.163 It is used by the
ecosystem. Indeed, many of the government’s AI Australian Federal Police (AFP) and the Department
initiatives may be too dispersed – and conceptual of Home Affairs to ethically research the use of
– to help organisations make sense of the new machine learning and data analytics in advancing
dynamics created by AI technologies. The Australian law enforcement.
Institute for Machine Learning (AIML) called for
the creation of a National Centre of Excellence in Some state governments, such as Queensland164
Machine Learning, an innovation hub that would and South Australia,165 have also declared their
bring together Australia’s AI research groups and intention to develop AI capabilities, often within
industry to meet the opportunities and challenges of innovation precincts. This is an area in which
AI technologies.162 there could be significant growth over the coming
years, although it could also be undermined as
Strengths state capitals compete for AI talent. Government
efforts have been relatively small thus far,166 but
The Australian government is focused on promoting Australia already has a number of centres of AI
social and inclusive growth and having the right

research, including Data61 and the 3A Institute (at country a competitive advantage. Titled “Developing
the Australian National University), along with noted Standards for Artificial Intelligence: Hearing
global experts in particular areas. The Westmead Australia’s Voice”, the resulting paper notes that
Applied Research Centre at the University of Sydney standards developed and applied specifically to AI
won the Google.org AI Impact Challenge to develop can not only guarantee AI is used for social good in
an AI programme that helps prevent heart attacks by Australia, they can also help other countries adopt
using data from smartphones and wearables.167 a like-minded human-centric approach to AI design
and deployment.173
Several Australian institutions are sharing AUD50
million (USD35million) from the Microsoft AI for In July 2019, the Australian Council of Learned
Earth programme to conduct research on water, Academies (ACOLA) launched “The Effective and
agriculture, biodiversity, and climate change.168 The Ethical Development of Artificial Intelligence – An
Federal Court developed a Proof of Concept using Opportunity to Improve our Well-being”, a paper
the Watson platform with IBM partner Carrington that calls on the government to lead efforts to make
Associates for splitting assets after a relationship Australia a global example for the responsible
breakdown.169 And National Australia Bank (NAB) adoption of AI.174 ACOLA concludes that the best
set up a Data Guild to educate bank employees in way to ensure AI contributes to Australians’ well-
the value of good data governance, data quality and being is to establish an independent policy and
ethics, working with AWS, Azure, and Google.170 regulatory body dedicated to the sole advancement
of AI. In August 2019, the Australian National
A budding legal and ethical framework for AI University (ANU) launched the AUD1.5 million (USD1
million) Humanising Machine Intelligence (HMI)
Since the beginning of 2019, Australia has multiplied project to build a framework for moral machine
the number of initiatives designed to ensure that intelligence (MMI) and develop an “AI with moral
AI innovation is done in an ethical and responsible values”.175 The initiative gathers some of the
manner. In January 2019, the Australian Human country’s foremost AI specialists from all levels and
Rights Commission shared the findings of a study branches of the AI spectrum, an achievement in and
on the way AI governance and leadership can of itself.
impact human rights.171 Conducted in conjunction
with the World Economic Forum (WEF), the project For Sassoon Grigorian, Senior Director of
highlighted the fact that most AI discussions focus Government Affairs & Public Policy at Salesforce,
on the right to privacy (who owns, controls, and all these initiatives are important for Australia to
exploits personal data), but there are many other catch up – and keep up – with AI development
areas in which AI can profoundly impact socio- in other countries. He notes in a blog post that
political realities (employment, policy-making, socio- these efforts could culminate in the creation of a
economic equality, etc.). National Advisory Council on ethics and AI, with
“members representing a mix of sectors including
In April 2019, the Department of Industry, Innovation business, not-for-profit, academia, and government,
and Science kicked off a national conversation on and a diverse range of backgrounds that reflect
how the country should develop and use AI by first community diversity. It should also not be limited to
identifying key governance principles and measures, technologists, including human rights advocates,
with the well-being of Australians as the top priority. ethicists, economists, and community members.”176
The discussion resulted in the “Artificial Intelligence:
Australia’s Ethics Framework” paper, released on
what should be the core AI principles and values
and whether they resonated with the Australian
public.172 In June 2019, Standards Australia began a
consultation process with key stakeholders across
industry, government, civil society, and academia on
standards for AI applications which could give the

Challenges
Lack of a dedicated national AI strategy
While the government has devised a range of initiatives conducive to the development of
AI under the broad policy agenda of digital economy and innovation strategy, they remain
fragmented across various agencies and initiative champions. As Australia has already discovered
through the development of its digital transformation journey, it will be beneficial for the
government to develop a comprehensive framework with strong strategic focuses for AI and
coordinate the execution of the national strategy to collaborate across different agencies to
implement the action plans.177
Gaps in AI talent and skill development
While venture capital is beginning to make itself available, and AI experts continue to call for
increased government spending (so as to keep up with other countries), the underlying challenge
lies in developing the necessary human capital and overcoming the skills gaps that currently
abound in both public and private sectors. This requires a long-term commitment that has been
slow to emerge – partly due to a climate of fear over automation, the ‘future of work’ and a
dismantling of privacy rights. Half of all Australian respondents to a Deloitte survey expressed
major or extreme concern with existing AI risks.178
Low business and consumer adoption
According to the Salesforce AI Readiness Index, Australia scores well in terms of government
readiness, but far less well in terms of business and consumer readiness.179 This discrepancy
between government effort and wider adoption risks turning Australia into what Tufts University
calls a “stall out” economy: one that has attained digital maturity in key areas, but whose overall
pace of innovation is slowing down. If consumer and business readiness do not improve, Australia
will be overtaken by nimbler, more dynamic economies that are eager to keep up with the rapidly
growing and evolving AI sector. Catching up will become harder the more advanced, complex,
and widespread AI technologies become.

China National AI Strategy Yes
Ministry of Science and Technology
National Agency
(MOST); AI Plan Promotion Office
Baidu, Alibaba, Tencent, iFlytek,
SenseTime, Ubtech Robotics, Cambricon
E-commerce, Manufacturing, Smart City,

Key Industry/Sector
Vehicles, Surveillance/Police, Finance180
Facial Recognition, Language/Voice
Key AI Technology Recognition, Robotics, Neural Networks,
Computer Vision
Rank in EIU’s Automation
12 / 25
Readiness Index181
Government AI Readiness 20 / 194
Index182
Three-Year
• Intelligent and networked
Action Plan to
products including
Promote the
connected vehicles, The Action plan lays out
Development of a
Ministry of Industry and service robots, video image strategies to achieve the
New-Generation
Information Technology identification systems first milestone in the Next
Artificial
(MIIT) • Support systems including Generation AI Development
Intelligence
sensors and neural network Plan
Industry (2018-
processing chips
2020) (December
• Intelligent manufacturing
2017)
National Development
AI is identified as one of
and Reform Commission • Smart home and intelligent
the 11 priority areas to help
(NDRC), the Ministry of appliances
accelerate the use of ICT in
Internet Plus Science and Technology • Intelligent unmanned
conventional industries. It
Artificial (MOST), the Ministry of systems
targeted to build platforms
Intelligence Plan Industry and Information • Smart automobiles
for fundamental AI innovation
(2016-2018) Technology (MIIT), • Robots
and make breakthroughs
and the Cyberspace • Intelligent wearable
on basic core technology by
Administration of China devices
2018.
(CAC)
MOST together with

the AI Development
The Development Plan
Plan Promotion Office • Autonomous vehicles
Next Generation comprehensively outlines
leads and engage • Smart cities
Artificial China’s strategies and
with 15 ministries • Computer vision in medical
Intelligence initiatives to accelerate
and departments diagnosis such as medical
Development economic and social
including the NDRC, imaging
Plan (July 2017) development through
MIIT to implement • Voice intelligence
advancing AI technologies.
and coordinate key AI
projects

• Create a conducive environment by investing in
industry training resources, standard testing, and
cybersecurity.
The top down approach adopted by the Chinese

government in driving R&D and commercialisation
of AI technologies has ensured IT companies and
start-ups have access to stable funding and financial
support. As of June 2018, there were more than 1,000
AI companies (start-ups and bigger organisations
combined) in China, as compared to half that many
in the United States.187 In 2018, 27 out of the top 50
Chinese AI companies were backed by either govern
ment-associated funds or the leading tech giants of
Country Overview Baidu, Alibaba, and Tencent.188
China has a national AI policy that clearly outlines its In addition to government funding, Chinese AI start-
strategy to become the world’s leading AI power by ups have been very successful in attracting venture
2030. Launched in July 2017, the Next Generation capital, securing 48% of total global investment
Artificial Intelligence Development Plan includes in 2017.189 Central government guidance funds
initiatives and goals for R&D, industrialisation, talent are established and channelled into technology
development, education and skills acquisition, development zones and science parks. Provincial and
standard setting and regulations, ethical norms, and city governments are also following suit to mobilise
security.183 resources and funding to bolster AI development,
promote collaborations among government, industry,
Before the Next Generation Plan, China had specific and academia with supporting policies.
AI-related provisions within its broader national
plans. The Made in China 2025 Blueprint184 and There is also the fact that Chinese citizens seem
the Three-Year Guidance for Internet Plus Artificial to be embracing AI with enthusiasm, using AI face
Intelligence Plan (2016-2018),185 for instance, recognition for everything from credit applications
outline the expansion of the robotics, autonomous to payment authentication. Unlike other economies,
vehicles, and smart homes industries. But the China’s citizens are more willing to adopt technology
Next Generation Plan marks a turning point, as it first, rather than wait for related regulations.190 This
made AI a clear national priority backed by central unparalleled access to large quantities of consumer
government guidance and public-sector funding. data generated from a huge population actively
using mobile devices and digital platforms has
Since the release of the Next Generation Plan, given Chinese AI companies a distinct comparative
the government has published the Three-Year advantage in research and development which
Action Plan to Promote the Development of New- involves the testing of algorithms and machine
Generation Artificial Intelligence Industry,186 which learning.
advances four major tasks ahead of the 2020
targets: Cited in a report, Dr Kai-Fu Lee, Chairman and CEO
of Sinovation Ventures and President of Sinovation
• Develop intelligent and networked products such Venture’s Artificial Intelligence Institute, notes that
as vehicles, service robots, and identification “the Chinese entrepreneurial system is an advantage
systems; for China. AI companies are moving very quickly into
• Enable AI support systems, including intelligent new spaces. A third big factor is the government
sensors and neural network chips; support. At a local level, cities give subsidies to
• Support the growth of intelligent manufacturing AI companies that move there, they have venture
systems and mechanisms; and capitalists that invest in AI, and they have smart

people move to these cities, including overseas- regulations, data sharing and the development of
returning experts.”191 China has already emerged ethical standards, Shanghai will concentrate their
as a credible AI global leader, accelerated by a efforts on AI applications in healthcare, smart cities,
powerful technology industry, a ‘mobile first’ society, intelligent transport.
and an open approach to data collection. Chinese AI
companies have a clear advantage in terms of both Leveraging cross-sector and global collaborations
funding and government support, with public and
private sectors working in tandem to achieve clear AI companies in China are highly innovative and a
and identified, albeit ambitious, goals. number of them have successfully teamed up with
internationally renowned research institutes and global
Strengths players, to acquire and advance the latest AI know-
how. For example, Baidu further strengthened its ties
Whole-of-government endeavour with strong with Intel to enhance the neural network processor
funding support and optimise its various AI platforms including deep-
learning, voice assistance and autonomous driving.196
China’s AI industry is booming: in 2017, it was valued Tencent’s Medical AI Lab, currently focused on the
at CNY23.7 billion (USD3.5 billion) growing by 67% diagnosis of Parkinson’s disease, is partnering with
that same year. China has consistently surpassed UK-based digital health tech start-up Medopad and
the United States in terms of patents and research specialists in King’s College London to assess and
papers since 2006,192 with 25% of AI papers on track patients’ movement using video analysis and
Scopus in 2017 originating from China as compared motion capture technologies.197
to 17% from the United States.193 Extensive venture
capital and government funding have contributed International collaborations and research networks
to the proliferation of both AI research and AI play a pivotal role in advancing AI technologies
companies: as of June 2018, there were more and will continue to drive breakthroughs in AI
than 1,000 AI companies (start-ups and bigger applications. In Shanghai, an AI Industry Alliance
organisations combined) in China, as compared was established to enhance resources matchmaking
to perhaps half that many in the United States. A among the participating enterprises to accelerate the
majority of these companies are located in Beijing, development of AI applications in different sectors.198
Shanghai, Shenzhen, and Hangzhou, specialising in The 22-member bloc includes BAT, DeepBlue, and
areas such as voice recognition, computer vision, UCloud, along with the Microsoft Asia Research Centre,
and natural language processing. and ABB, a Swiss conglomerate specialising in robotics
and automation technology.
Central government guidance funds are established
and channelled into technology development Nascent AI principles and regulatory frameworks
zones, robotics, and science parks. Provincial and
city governments are also following suit to mobilise A coalition, comprising the Beijing Academy of Artificial
resources and funding to bolster AI development, Intelligence (BAAI), Peking University, Tsinghua
promote collaborations among government, University, Institute of Automation and Institute of
industry, and academia with supporting policies. Computing Technology within the Chinese Academy
Shanghai alone will set up investment funds of no of Sciences, and the BAT, released the Beijing AI
less than CNY100 billion (USD15.1 billion)194 and the principles in May 2019 to provide guidance on AI R&D,
Tianjin government has contributed to establishing implementation, governance and long term planning.199
a CNY100-billion industrial fund.195 Beijing and These principles stress the importance of respecting
Shanghai set up pilot zones for AI in February and people’s privacy, freedom, dignity, rights and autonomy
May 2019 respectively after an earlier commitment and are largely in line with those adopted by the OECD
to build an AI-centric development park in the and the EU.
capital city that could host 400 enterprises. While
the Beijing pilot zone will focus on AI-related

This formulation of AI principles is one step closer can be expected to accelerate the development of
to the establishment of comprehensive ethical standards and legal frameworks governing issues
and regulatory frameworks, and policies which the such as the use of robotic systems, computer
country targets to achieve by 2030. In view of the vision, human-machine interaction, AI security, and
rapid development in AI, the Chinese government accountability in cases of AI-related accidents.

Challenges
Domestic talent shortage
With AI high on the agenda of many economies, competition for AI talent is rife and there is a
global shortage of AI researchers. Although China had the world’s second largest AI talent pool
next to the United States in 2017, accounting for 8.9% of the world’s total, only 5% are top-rated AI
specialists with many years of research experience.200 Municipal governments, research institutes
and IT companies alike have been relying on short term strategies such as scholarships and
competitive remuneration packages to attract talent from overseas and retain local talent.201
To bridge the talent gap, the government needs to devote more resources to universities to
boost the number of undergraduate and graduate students taking not only core STEM subjects,
but various creatively challenging subjects that will be necessary for holistic and advance AI
development, as well as further nurturing AI talent through exchange and training programmes. As
industry plays a critical role in cultivating talent, industry–university collaboration can expect to be
fostered where industries can assist in setting the curriculum and provide students with internships
and the opportunities to work in AI innovation labs.
Putting business interests ahead of research
AI Investment from China’s tech giants including Baidu, Alibaba, Tencent (BAT) have been skewed
towards the development of applications in a comparatively small number of select sectors such as
healthcare, finance, and connected vehicles. Contrastingly, relatively fewer resources have been
devoted to AI research (as well as the core technologies that underpin the development of AI).
This over-emphasis in developing profit-driven AI applications not only intensifies competition

among start-ups, but also risks leading to over-investment and duplication of resources. Research
can be strengthened through an integration of industry and academic R&D to strike a balance
between applications development and fundamental research. Home-grown core technologies
may become critical competitive enablers moving forward, particularly if various trade disputes and
global value chain fragmentation continues.

Indonesia National AI Strategy No
Ministry of Communication and

National Agency
Information Technology (Kominfo)
Tokopedia, Snapcart, Bukalapak,

Key AI Organisations Nodeflux, Kata.ai, Bahasa.ai, Dattabot,
Prosai.ai, Eureka.ai
E-commerce, Manufacturing, Smart City,

Key Industry/Sector Vehicles, Education, Advertising/Customer
Experience, Cybersecurity202
Language/Vocal Recognition, Machine

Key AI Technology Learning, Recommendation Engines,
Chatbots, Intelligent Teaching Modules

25 / 25
Readiness Index203

Index204
GR 71/2019 was introduced in October

Government
2019 to replace GR82/2012. It redefines
Regulation
key concepts such as “Personal Data”,
No 71 of 2019
introduces new differentiators such
(GR 71/2019) Government-
Cross-sectoral as private and public categories for
on Electronic wide
electronic system organisers (“ESOs”), and
Systems and
widens the scope of government action
Electronic
when governing electronic systems and
Transactions205
platforms.
Five key technological

advances:
Revitalise the manufacturing sectors in
• internet of things (IoT), a bid to increase the contribution of to
Making
Ministry of • AI, GDP growth and elevate Indonesia to a
Indonesia 4.0
Industry • Human-machine global top 10 economy by 2030 through
(2018)
interface, expanding net exports, and enhancing
• Robot and sensor productivity and innovation capabilities.
technology,
• 3D printing.

GR 95/2018 was introduced in 2018, and

governs the Sistem Pemerintahan Berbasis
Elektronik (SPBE), which is the government
administration that utilises information
and communication technology to provide
services.
Presidential
This includes aspects of the country’s
Regulation No Government-
Cross-sectoral government cloud, including the SPBE
95 of 2018 (GR wide
infrastructure, defined as hardware,
92/2018)
software, and facilities that become the
main support for running the systems,
applications, data communications, data
storage and processing, integration/
connecting devices, and other electronic
devices, and also the country’s National
Data Centre.
Produce high-quality digital technology

Priority sectors:
start-ups valued at USD10 billion, through
1,000 Start-Up incubation and cooperation between
• Agriculture
Digital National Ministry of academia and start-ups.
• Education
Movement Communications
• Health
Programme and Information It will be launched in 10 cities including
• Tourism
(2016) Jakarta, Bandung, Surabaya, Yogyakarta,
• Logistics
Semarang, Malang, Medan, Bali, Makassar,
• Energy
and Pontianak.
Country Overview
Indonesia is very open to AI, but it does not

have a comprehensive national AI strategy to
effectively frame this openness. It does have several
nationwide plans that aim to turn the country into the
largest digital economy in Southeast Asia by 2030,
but none of them target AI specifically.
In April 2018, President Joko “Jokowi” Widodo

launched the Making Indonesia 4.0 programme
to help the country better compete in the digital
era.206 A comprehensive roadmap to digitisation, the
programme contains ten steps that the government
needs to implement, including: empowering SMEs
through digital technology, improving digital

infrastructure, attracting foreign investors, changing the Indonesian market “influenced by trends in
educational curricula, and providing incentives the global market and flooded by a wave of new
for businesses based on digital technology. Most technologies that might not be truly relevant or
importantly, the programme explicitly mentions viable to be implemented straight away. Listening
five key technologies that Indonesia must prioritise to customers prior to introducing a new product,
in the years to come: Internet of Things (IoT), AI, service, or technology is always recommended as
Human Machine Interfaces (HMI), Robotics and the best practice.”211
sensor technologies, and 3D-Printing technologies.
Together, these trends demonstrate that Indonesia is
Indonesia’s start-up ecosystem is growing rapidly, still in the initial stages of AI development, especially
as evidenced by the number of success stories in comparison to neighbouring countries. And this is
coming out of the country. Ride-hailing firm Go- despite the recognition among key enterprises and
Jek, web forum Kaskus, e-commerce platform certain government agencies of the importance of
Tokopedia, and most Fintech players, are very leveraging the opportunities offered by advanced
successful in their respective fields, and are actively AI technologies such as automation, robotics, and
adopting AI solutions to further consolidate their autonomous vehicles.212
competitive advantages. A key differentiator for
these companies is the trove of data they are Dr Nyoman Adhiarna notes that turning Indonesia’s
accumulating to build targeted AI models and openness to AI into effective, sustainable adoption
applications. With a population size of over 267 will require a consistent and concerted approach:
million, Indonesia (like China) can stand to leverage “From a regulatory perspective, there is no
large amounts of user data that will power the specific law related to AI. There are the Electronic
sustainable advancement and spread of AI.207 Information and Transaction Law, the GR82 [now
GR71] regulation in relation to data protection
The Indonesian government is also driving the (principles of transparency and accountability), and
country’s AI readiness. According to Dr Nyoman the future Personal Data Protection Law (currently
Adhiarna, IT Governance for Electronic Systems and being drafted) – but nothing to guide institutions
Digital Economy, Ministry of Communication and and organisations on the unique ins and outs of
Information Technology (Kominfo), key government AI. Likewise, putting a specific government agency
agencies are proactively launching AI-driven in charge of AI initiatives could help consolidate
initiatives in a range of sectors: “The Central Bank current and future AI efforts. Kominfo could, for
and the Tax Department are using AI to analyse their instance, supervise and coordinate the AI initiatives
own data and improve their policy-making decisions. launched across different economic sectors.”213
The Ministry of Health is looking into AI to transform
patient services as well as administrative processes. Strengths
The Agency for Climatology, Meteorology, and
Geophysics (BMKG), meanwhile, is preparing to A young and dynamic start-up ecosystem
leverage the weather and earthquake data it owns,
though for now data ownership is fragmented A number of AI companies have emerged in
across multiple agencies.”208 Jakarta in the last few years with the potential to
thrive and to drive the focus on, and development
On paper, then, Indonesia is particularly open to of, AI. For example, Nodeflux, a Jakarta-based AI
AI.209 An IDC survey found that at least a quarter of start-up backed by Telekom Indonesia and venture
Indonesian technology companies are adopting AI, capital East Ventures, developed expertise in facial
with the country leading the region in this regard.210 recognition technology, computer vision and video
But, it is argued, many of them may be adopting AI analytics that can be employed in people counting,
solutions that are not necessarily relevant or useful crowd and traffic management, and monitoring of
to actual business needs. Edwin Sugianto, COO illegal parking. The company also partnered with
of Asuransi Axa Indonesia, is quoted as seeing the provincial government in Jakarta to detect

water levels in dams, sewers and waterways and other e-commerce giant Tokopedia partnered with
promptly issues warnings and alerts when flooding Universitas Indonesia to create the country’s first
is anticipated. Nodeflux plans to further solicit AI Centre of Excellence with support from Nvidia.
government support to build an AI community in Leveraging Nvida’s deep learning supercomputer
Indonesia. technology, researchers can develop and test a wide
range of AI solutions tailored for different industries
Other AI start-ups such as BJtech, Halosis and Kata. such as logistics, transport, and financial services, in
ai, which partnered with mobile network operator addition to solving societal issues.217
Telkomsel, have been successful in building
conversational AI platforms using natural language
processing technologies. Together with the country’s
two major e-commerce platforms, Tokopedia and
Bukalapak, these AI start-ups have accumulated
a trove of data that can be used to further build AI
models and applications. With a population size of
over 267 million, of which some 106 million are tech-
savvy mobile Internet users, Indonesia (like China)
stands able to leverage large amounts of user data
that is of paramount importance in fuelling the
development of AI.214
A strong potential for AI adoption among

enterprises and SMEs
A survey conducted with businesses engaged in the

retail, IT/telecom, financial services and insurance
industries across eight APAC markets by Appier
found that Indonesian companies top the list in AI
implementation with 65% adopting or expanding
the use of AI to improve operational efficiency and
develop innovative products and solutions that
better meet customers’ needs. The results suggest
a high level of willingness to explore and utilise AI.215
In this regard, helping SMEs acquire a broad base of
AI awareness and acceptance can further enhance
adoption, allowing them to take advantage of these
emerging technologies and encourage both greater
usage and further innovation from AI start-ups.
Innovative public-private partnerships
In February 2019, e-commerce platform Bukalapak

partnered with the Bandung Institute of Technology
to inaugurate the country’s first AI and cloud
computing innovation centre. 200 engineers from
Bukalapak are to engage in knowledge exchange
with researchers and students. More importantly,
the company will contribute big data to facilitate
research on natural language processing, computer
vision and deep learning.216 In March 2019, the

Challenges
By 2021, more than 30% of consumers in Indonesia

are expected to interact with AI-enabled bots and
one-fifth of commercial enterprise applications are
set to be powered by AI.218 However, compared to
neighbouring countries, Indonesia is still in the initial
stages of AI development, despite the recognition
30%
among key enterprises and certain government
agencies of the importance of moving from
labour-intensive to knowledge-intensive growth
models and leveraging the opportunities offered
by advanced AI technologies such as automation,
robotic systems, autonomous vehicles.219 of consumers in Indonesia are expected
to interact with AI-enabled bots
“It is increasingly difficult to

find Indonesian talent that
has the experience to take
[AI] to the next level.”
Aswin Tanu Utomo, Vice President of
Engineering at Tokopedia

Lack of a national AI strategic plan
Although AI and robotic sensor technologies are key focus areas within the Making Indonesia 4.0
roadmap, there are no specific strategies and action plans to shape the direction and development
of AI – even though it is expected to be one of the key enablers for Industry 4.0. The government
will need to formulate a national plan with coordinated strategies across relevant agencies if it is
to accelerate research, innovation and adoption to capitalise on the envisioned AI opportunities. A
dedicated government agency could be set up to identify key sectors to be focused on, examine
the policy and ethical implications of AI deployment, determine the roles of different ministries and
departments, and coordinate work among them.
A small pool of AI talent
Currently there is a very limited AI talent pool for Indonesian companies to tap into, and the country
has a relatively small proportion of students studying STEM subjects in top-rated universities. Cited
in an online article, Aswin Tanu Utomo, Vice President of Engineering at Tokopedia, believes in
Indonesian talent, but he also recognises that “It is increasingly difficult to find Indonesian talent
that has the experience to take [AI] to the next level.”221 The government has proposed to increase
the overall education budget to IDR505.8 trillion (USD35.5 billion) in 2020, with part of that sum
devoted to developing domestic AI capabilities via college scholarships and AI skills-training
(coding, data analytics, etc.).222
Infrastructure often does not enable continuous/sustainable connectivity
In Indonesia, cloud infrastructure and data centre facilities are limited, and telecommunication
infrastructures are still underdeveloped. For example, fixed and mobile broadband speeds in
Indonesia are one of the lowest in ASEAN behind Laos, Philippines, and Vietnam220 and many
rural parts of the country are still deprived of high-speed mobile networks. There is also the fact
that many rural/remote parts of the country are both susceptible and vulnerable to major natural
catastrophes, jeopardising what little connectivity these areas may have. Mobile network and
fibre infrastructure will go a long way in connecting remote areas and islands, improving internet
accessibility and broadband speeds. Developing a national data centre and a cloud network
may also help overcome this obstacle.

Malaysia National AI Strategy Yes (forthcoming, announced in 2018)
Malaysia Digital Economy Corporation
National Agency
(MDEC)
G3 Global, NEUON, Avanade, Glueck

Technologies, Wonderland Technologies
E-commerce, Manufacturing, Agriculture,

Key Industry/Sector Smart City, Surveillance/Police, Financial
Services
Facial Recognition, Robotics/Robo-Advisors,

Key AI Technology Machine Learning, Sentiment Analysis,
Chatbots

14 / 25
Readiness Index223

Index224
In October 2019, Budget 2020 was

announced, with a clear intention to
drive the digital transformation of society
Government-wide and institutions alike. A number of
Budget 2020 Cross-
(led by the Ministry of measures were announced to support
(2019)225 sectoral
Finance) start-ups, prepare 5G deployment,
train the workforce, and accelerate
the adoption rate of next-generation
technologies (including AI).
National AI Framework Malaysia Digital Drive Malaysia’s AI ecosystem, by

Cross-
(forthcoming, announced in Economy leveraging AI and machine learning for
sectoral
2018)226 Corporation (MDEC) digital transformation
Aims to drive the country towards

Cross-
becoming an Industry 4.0 hub for
National Policy on Industry Ministry of sectoral,
Southeast Asia, using new technologies
4.0 (Industry4WRD) International Trade with
such as IoT, AI, Robotics, and Additive
(2018)227 and Industry emphasis
Manufacturing to increase productivity
on industry
and competitiveness across industries.
Drive big data analytics (BDA) demand

National Big Data Analytics Cross-
MDEC in all sectors and activate adoption &
Framework (2015)228 sectoral
usage of BDA in the public sector.

“[Malaysia’s National AI Framework] will
develop strategies and action plans
specifically for AI. Its main objective will
be to enable a dynamic AI ecosystem
and to leverage the social and
economic benefits that AI technologies
can potentially deliver.”
Eric Ho, Investment Research & Advisory at MDEC
Country Overview
The Malaysian government is finalising a National • Security and surveillance: Since 2017, MIMOS,
AI Framework, expected to be completed by the Malaysia’s national research and development
end of 2019. It will be led by the Malaysia Digital centre for ICT, is working with Huawei to develop
Economy Corporation (MDEC) through a newly AI-based security and surveillance solutions,
formed AI unit,229 which will comprise both local and including advanced video analytics and facial
international experts to ensure that Malaysia is on recognition systems.234
the right track to develop a robust AI ecosystem.230
According to Eric Ho, Investment Research & For Dr Norisma Idris, Associate Professor at the
Advisory at MDEC, the National AI Framework will Department of Artificial Intelligence (University
“develop strategies and action plans specifically of Malaya), “It is important that the government,
for AI. Its main objective will be to enable a industry, and universities continue to work together
dynamic AI ecosystem and to leverage the social to support the national AI agenda. In many countries,
and economic benefits that AI technologies can these partnerships enable the development and
potentially deliver.”231 In the meantime, Malaysia’s AI commercialisation of innovative AI products and
strategy is strongly based on fostering enhanced services, and Malaysia has everything it needs to
cooperation mechanisms between government, foster a dynamic, homegrown AI ecosystem.”235
academia, and industry, as demonstrated by several
strategic partnerships launched in the smart city and In the last few years, Malaysia’s start-up and
e-commerce sectors: technology ecosystem has indeed undergone a
massive transformation. The start-up community has
• The Malaysia City Brain project: A partnership attracted investments to the tune of USD1.45 billion
between MDEC, Kuala Lumpur City Hall, and in 2016,236 and incubated a number of deep-learning
Alibaba, the initiative combines 5G, IoT, and AI start-ups that are using AI for e-commerce activities,
technologies to optimise traffic, parking, and human sentiment analysis, and automated customer
energy management.232 support, among others.
• The digital free trade zone (DFTZ): Alibaba also
worked with MDEC to set up the DFTZ – an For Dr Norisma Idris, sentiment analysis – the
eFulfillment and eServices hub designed to use of natural language processing, text analysis,
facilitate cross-border trade and enable local computational linguistics, and biometrics to identify,
online businesses to export their goods.233 extract, quantify, and study affective states and

subjective information – has become a key contains an array of personal health management
differentiator for businesses in Malaysia: “Sentiment options including a dengue alert feature called
analysis is the main AI trend in Malaysia. Most Artificial Intelligence in Medical Epidemiology (AIME),
industries and many government agencies perform which utilises big data and AI to predict whether a
sentiment analysis for specific purposes. certain locality will face a dengue epidemic within
This includes feedback from customers (for the next 30 days, with an accuracy of 80%.241
industry), and the prediction of positive/
negative perceptions on key policy issues (for Likewise, Plus, Malaysia’s largest highway
government).”237 concessionaries company, is partnering with
Microsoft Malaysia to integrate the Microsoft
Broadly speaking, the private sector has been very Azure cloud system into its highway toll network
active in developing innovative uses of AI. Primarily to transform and modernise the country’s highway
led by international organisations, these initiatives experience. This partnership would make it the first
are being developed in various sectors, including in the country to leverage AI, big data analytics and
transport, logistics, and energy. Dr Norisma Idris cloud computing to improve its highway network by
notes that “Just a few years ago, the medical better predicting traffic congestion and monitoring
industry was driving AI the fastest, as advanced the safety of its highways.242 Alibaba Cloud, aside
image-processing solutions were being developed from supporting the government’s digital agenda
for all types of medical devices. Today, however, in Kuala Lumpur through the City Brain initiative
language-driven applications such as chatbots, to monitor and collect data on traffic movement to
sentiment analysis, and text normalisation systems improve mobility and alleviate congestion,243 is also
dominate in most industries.”238 working with Bank Muamalat to use big data and AI
to better understand and develop tailored products
Despite these factors, businesses and and services for different customer segments.244
organisations in Malaysia remain unprepared to
adopt AI; organisations have yet to develop or This openness with working with international
articulate a clear AI roadmap. According to an experts with greater expertise and experience
IDC survey, only 26% of Malaysian organisations in various sectors not only enables technology
have embarked on their AI journeys. As it stands, transfers and knowledge sharing, but also helps to
the country’s adoption rate (8.1%) lags significantly build awareness and encourage other organisations
behind other APAC markets,239 highlighting the to follow-suit.
urgent need for a consolidated AI strategy that is
both driven and owned by Malaysia. As Eric Ho AI park to foster R&D and grooming local talent
of MDEC puts it, “AI development and adoption
can be accelerated in Malaysia by focusing The Malaysian company G3 Global has inked a
on constructing comprehensive polices and USD 1 billion-agreement with two Chinese firms
regulations for better direction and vision.”240 to develop its first AI park, focused on building
AI applications that take advantage of computer
Strengths vision, speech recognition and natural language
processing. The park will also serve as a platform
Robust private sector development and to groom local AI talent, and support efforts to
international collaborations build a commercial AI ecosystem and advance AI
research in Malaysia. This park is expected to have
In the absence of a national AI strategy, the private a huge impact on government agencies, as well as
sector has been developing innovative uses of banking, manufacturing and healthcare industries
AI. Primarily led by international organisations, by centralising R&D akin to a mini ‘Silicon Valley’
these initiatives have been developed in various for AI in Malaysia.245 Further, this would also help
sectors. For example in health, Prudential Malaysia to retain local AI talent by developing an active
launched “Pulse”, an all-in-one application that hub for innovation and development to realise
offers AI-powered health information which opportunities.

In the last few years, Malaysia’s start-up and and automated customer support among other
technology ecosystem has undergone a massive things.246 Concurrently, AI centres of excellence and
transformation, with the start-up community consultancies such as Avanade247 and Crayon248
attracting investments to the tune of USD1.45 are offering a range of services to assist companies
billion in 2016, according to a report by Pagan in effectively implementing AI in their business
Research. The Malaysian start-up ecosystem has operations. However, this continues to be an area
also incubated a number of deep tech start-ups the government can do more in to simulate domestic
that are using AI for detecting fraudulent behaviour talent development, especially given the gap in local
in e-commerce activities, human emotion analysis human resources.
Challenges
Lack of open discussions with stakeholders
Where the National AI Framework was first announced by the previous Najib-government, there
were question marks on the continuity of the Framework. This was put to rest and the new
government has announced a target of end-2019 for the National AI Framework. However aside
from couple of media announcements that it is being developed, there is little detail released.
Alongside the National AI Framework, the “Cloud First” Strategy meant to guide public sector on
how to adopt cloud computing technologies has yet to be released, while the Digital Trusted Zone
has also seen little progress.
Without a transparent process involving multi-stakeholders including local industry players, it is

unlikely that the National AI Framework will have incorporated the ambitions or concerns from
local stakeholders which may lead to an implementation gap.

Singapore National AI Strategy Yes
AI Singapore, National AI Office,
National Agency Infocomm Media Development Authority
(IMDA)
Key AI Organisations Trax, CXA, Rotimatic, Zhixin, Evie
E-commerce, Smart City, Vehicles,

Key Industry/Sector Surveillance, Financial Services,
Healthcare, Education
Facial/Vocal Recognition, Robotics,

Customer Experience, Chatbots, Robo-
Key AI Technology Advisors, Machine/Deep Learning,
Predictive Analytics, Automated
Algorithms
3 / 25
Readiness Index249
Index250
Long-term national programme to

accelerate the development and adoption
of AI across Singapore society, starting
National AI Strategy with nine key sectors: Transport and
National AI Office Cross-sectoral
(2019)251 Logistics, Smart Cities and Estates, Safety
and Security, Healthcare, Education,
Manufacturing, Finance, Cybersecurity,
and Government.
In November 2019, the MAS announced

Financial
Veritas Monetary Authority of Veritas, a framework for financial
services
(2019)252 Singapore (MAS) institutions to promote the responsible
industry
adoption of AI and data analytics.
Provide an accountability-based
Model Artificial
framework with ethical principles for
Intelligence (AI) Personal Data
organisations to practically implement
Governance Protection Commission Cross-sectoral
when deploying AI – ultimately promoting
Framework (PDPC)
AI adoption and building consumer
(2019)253
confidence and trust

Principles to Promote
Fairness, Ethics,
Accountability and
Outlining principles for the use of AI and
Transparency (FEAT) Financial
data analytics (AIDA) in decision-making
in the Use of Artificial MAS services
in the provision of financial products and
Intelligence and industry
services.
Data Analytics in
Singapore’s Financial
Sector (2018)254
To better uses data, connectivity and

computing to re-engineer business
processes, re-architect technology
Digital Government
Gov Tech Public sector infrastructure and transform services for
Blueprint (2018)255
citizens, businesses and public officers;
automates processes to better serve
citizens
Government-wide
partnership –
Infocomm Media
Development
Authority (IMDA),
Five-year, SGD150 million national
National Research
programme to enhance Singapore’s
Foundation (NRF),
AI Singapore (2017)256 Cross-sectoral capabilities in AI based on three pillars
Smart Nation and
of AI Research, AI Technology, and AI
Digital Government
Innovation.
Office, Economic
Development Board,
SGInnovate and
Integrated Health
Information Systems

Despite its small size, Singapore is home to 346
AI start-ups,263 compared to larger economies
such as China with 910264 and Australia with
278.265 Additionally, nearly 10% of organisations
in Singapore are already embracing AI in some
capacity, as highlighted in a 2019 Enterprise
Cognitive/AI survey.266
Singapore’s government-wide efforts are

largely focused on promoting a responsible and
accountable use of AI. To this end, the government
created the Advisory Council on the Ethical Use of
AI and Data,267 published several documents on
the ways responsible data collection and usage
can drive AI adoption,268 and launched a Model AI
Governance Framework.269 In addition, the Monetary
Authority of Singapore (MAS) released a set of
Country Overview fairness, ethics, accountability, and transparency
(FEAT) principles to promote the responsible use of
Singapore is one of a handful of APAC economies AI and data analytics in finance.270
to have all the elements in place to effectively drive
AI readiness and adoption.257 From data protection Apart from promoting industry AI adoption, the
laws258 to nationwide cybersecurity strategies,259 government is also driving more inclusive use of AI
Singapore has built strong regulatory foundations to in the public sector. Published in 2018, the Digital
maximise the impact of AI and other emerging digital Government Blueprint (DGB) was launched as a
technologies on its economy.260 component of Singapore’s Smart Nation initiative
focused on digital transformation of the government.
Singapore’s commitment to a progressive and The DGB sets a number of milestones for 14 specific
conducive approach to AI is embodied by AI initiatives to be achieved by 2023, including the goal
Singapore, a five-year national programme that all ministries and related government agencies
exclusively devoted to enhancing Singapore’s should have at least one AI project by 2023.271
AI capabilities through AI investment and AI
adoption.261 AI Singapore consists of four key In this context, policy-makers are keen to build
pillars, all aiming to drive investments and broaden the foundations for and drive the rapid take up
adoption:262 of AI, given its presumed potential to strengthen
Singapore’s economy – adding up to USD215 billion
• Fundamental AI Research: Funds scientific in gross value across 11 industries by 2035272 – and
research that will contribute to the other pillars of positioning Singapore as the data, algorithmic, and
AI Singapore. talent hub for the region.273
• Grand Challenge: Supports the work of multi-
disciplinary teams that provide innovative Strengths
solutions to major challenges Singapore and the
world face (health, mobility, finance, etc.). A strategy focused on accountable and
responsible development
• 100 Experiments: Funds scalable AI-based
solutions to industry-identified problems.
Singapore is focusing on managing community
• AI Apprenticeship: A nine-month structured
concerns around AI by grappling with issues of
programme to foster a new cohort of AI talent in
transparency, fairness, competition, privacy and
Singapore.
ethics in its policy and regulatory frameworks.

It is one of the few governments to create an AI (TR 68) were introduced by Enterprise Singapore
Ethics Advisory Council as part of its AI strategy, in January 2019 to promote the safe deployment
as well as provide guidelines and a framework for of autonomous AV in Singapore.281 Other initiatives
organisations to develop and use AI in a responsible to promote AV in Singapore include the National
manner. Promoting trust and confidence is a key University of Singapore (NUS) and the Singapore-
aspect of Singapore’s AI approach, and will be a MIT Alliance for Research and Technology (SMART),
useful example for other governments to emulate as the A*STAR’s Institute for Infocomm Research
they formulate their own AI strategies.274 (I2R),282 and the Centre of Excellence for Testing
Apart from the Model AI Governance Framework, and Research of Autonomous Vehicles (CETRAN)
initiatives announced in 2018 include the setting at Nanyang Technological University.283 In 2018,
up of an Advisory Council on the Ethical Use of the Ministry of Transport announced that the Civil
AI and Data, and a research programme on the Aviation Authority of Singapore (CAAS) is launching
Governance of AI and Data Use in partnership with an Aviation Transformation Programme (ATP) that
Singapore Management University.275 The Infocomm will promote the use of new technologies such as AI
Media Development Authority (IMDA) is also setting to improve airport operations.284
up a Digital Services Laboratory (DSL) programme to
bring together relevant parties to identify strategic Singapore also announced a collaboration with the
areas benefiting from faster intermediation, and World Economic Forum’s (WEF) Centre for the Fourth
address digitalisation challenges. Technologies Industrial Revolution (C4IR) to further drive AI and
or solutions created by IMDA such as AI libraries, data innovation, including engaging organisations
will be released by DSL as open source. Currently, to discuss Singapore’s Model AI Framework and
the National Speech Corpus and the Intelligent facilitate its adoption. The C4IR and IMDA will
Sensing Toolbox are available for download.276 The also develop a measurement matrix for the Model
Artificial Intelligence Technical Committee (AITC), Framework which regulators and certification
formed under the Information Technology Standards bodies globally can adopt for assessing whether
Committee (ITSC), will recommend the adoption of organisations are responsibly deploying AI. IMDA
relevant international AI standards for Singapore, and the C4IR will also develop a discussion paper
and further promote awareness of AI emerging outlining policy options on the facilitation of cross-
standards.277 border data flows.285
Human capacity development and private-public

collaboration
In terms of human capacity and skills development,

AI Singapore, launched in May 2017, is a five-year,
SGD50 million (USD36 million) government-wide
partnership to enhance Singapore’s capabilities
in AI.278 The programme consists of four key
initiatives focused on AI research, innovative
solutions, experiments and an AI apprenticeship
programme.279 AI Singapore launched two capacity
building programmes in 2018 – AI for Everyone
(AI4E) and AI for Industry (AI4I) – aimed at equipping
12,000 Singaporeans over three years with AI
knowledge.280
One key focus of AI development in Singapore is

on autonomous vehicles (AV). A set of provisional
national standards called Technical Reference 68

Challenges
Lack of soft skills and AI talent gap
Now, more than ever, soft skills, critical thinking, and knowledge of humanities (history, politics,
philosophy, sociology) are needed to make high tech products and services useful to society.
This is especially important in the case of AI, as citizens increasingly call for governments to
put ethical principles and values at the heart of AI plans and initiatives. The PDPC and MAS’
respective frameworks have begun looking at this, but are only scratching the surface where it
still remains unclear how to implement and apply these frameworks in AI.
On the talent side, this means building pools of talent that can in turn build and sustain digital
capabilities. Whether it is specialised technical talent (data scientists, cybersecurity experts,
machine-learning programmers) or digital transformation specialists (user-interface designers,
software engineers, business developers), it is important to be able to find the right people for
the right jobs. Singapore’s Smart Nation fellowship, for instance, was designed to attract and
develop technical talent that can innovate for the public good.286
Major cybersecurity lapses
The cyber-attack on SingHealth in 2018 saw 1.5 million personal particulars stolen. Investigations
revealed that SingHealth had not received the necessary security software updates for 14 months
since the last update on the spread of the WannaCry ransomware. Following this, and despite the
nationwide calls to tighten cybersecurity, another unauthorised access attempt was reported to
have taken place involving about 70 HealthHub accounts. And while the government continues
to ramp up resources to strengthen cybersecurity through its Cyber Security Law and the Cyber
Security Strategy,287 it needs to start at the root of the problem and identify current gaps in its own
infrastructure, be they technical, political, or operational.

South Korea Yes (forthcoming, announced in
National AI Strategy
2016)
National Agency Ministry of Science and ICT (MSIT)
Key AI Organisations Samsung, LG, Naver, Hyundai
E-commerce, Smart City, Vehicles,

Key Industry/Sector Financial Services, Education,
Healthcare

Key AI Technology
Chatbots, Robo-Advisors
Rank in EIU’s Automation Readiness

1 / 25
Index288
Rank in Oxford Insights’ Government

26 / 194
AI Readiness Index289
Act on Promotion
of Information and
In March 2019, several amendments were
Communications
Korea introduced to strengthen data protection
Network Utilisation
Communications Cross-sectoral standards and to align them with key
and Information
Commission GDPR requirements (a way to gain access
Protection (“Information
to the European market).
and Communications
Network Act”)290
Mid- to Long-Term This policy sets out the overarching

Master Plan in vision of a human-centred intelligent
Ministry of
Preparation for the information society, and proposes a three-
Science and ICT Cross-sectoral
Intelligent Information prong, balanced policy regime which
(MSIT)
Society has influenced and informed the AI R&D
(2017-2035)291 Strategy.
A joint effort by 21 government

Cross-sectoral departments, agencies and the Fourth
Industrial Revolution Committee, the
The policy seeks to milestone policy is focused on three main
promote intelligence, pillars (D-N-A): Data, Network, and AI.
innovation,
inclusiveness, and Going into its second year, there have
I-Korea 4.0 (2018-
MSIT interaction in a wide been several other policies and initiatives
2022)292
range of sectors introduced, by individual ministries as
including Smart well as collaborations between ministries,
Cities, Healthcare, to achieve the I-Korea 4.0 goals. These
Transportation, policies include the AI R&D Strategy,
Education, and the 6th National Informatisation Master
Manufacturing. Plan (2018-2022), and the Data and AI
Economy Revitalisation Plan (2019-2023).

Geared at developing three areas of AI:

technology, application to industry, and
AI R&D Strategy293 MSIT Cross-sectoral
support systems such as education and
employment.
Cross-sectoral
The policy articulates four goals that
are aligned with previous polices
Areas of focus:
released on intelligent technologies and
informatisation.
Smart Cities/
6th National
Villages/Factories,
Informatisation Master MSIT The four goals are: (1) an intelligent
Healthcare, Disaster/
Plan (2018-2022)294 and responsible digital government,
Environment
(2) economic growth through digital
Management,
innovation, (3) a human-oriented digital
Transport,
society, and (4) a secure and intelligent
Manufacturing and
network infrastructure.
Finance.
Cross-sectoral
The plan proposes nine policy tasks that
Initiatives include: promote data collection, distribution,
Data and AI Economy • Creating a and utilisation – ultimately raising the
Revitalisation Plan MSIT convergence domestic data market to KRW30 trillion
(2019-2023)295 cluster of 10 AI (USD24.7 billion); creating world-class AI
unicorn firms innovative ecosystems; and promoting AI-
• Training 10,000 data convergence.
data and AI talents
Country Overview
South Korea is energetically developing its AI

industry, seeking to become a top-four global
AI market and close the gap with leading AI
countries.296 With plans to finalise a national AI
strategy by the end of 2019, the South Korean
government is making considerable efforts to
accelerate domestic investment, adoption, and
innovation in AI.
In May 2018, as part of the national I-Korea 4.0 plan,

the Ministry of Science and ICT (MSIT) announced a
new AI R&D Strategy, bolstered by an investment of
KRW2.2 trillion (USD1.81 billion).297

tendency for government agencies to work in silos.
Comprising three main components, the strategy The MSIT is working with the MOTIE, the Ministry
aims to: of Food and Drug Safety (MFDS), and the Ministry
of Health and Welfare (MOHW) to invest KRW42
• Secure AI talent: Establish six graduate schools billion (USD34.6 million) over the next five years to
devoted to AI by 2022, with the goal of training accelerate the development of AI-powered medical
5,000 AI specialists (1,400 AI researchers and devices and systems.302 This pan-government
3,600 data management specialists); initiative is indicative of the government’s
• Develop AI technology: Fund large-scale projects recognition of the need to coordinate scarce – and
in national defence, medicine, and public safety scattered – resources to compete with nimbler
through an AI R&D challenge similar to DARPA in economies.
the United States;
• Support the development of AI start-ups and Strengths
SMEs: Support emerging AI businesses by
creating an AI-oriented start-up incubator. Local talent development, early ethics discussions,
industry collaboration
More recently, the Ministry of Trade, Industry and
Energy (MOTIE) indicated that it will introduce AI- As outlined in I-Korea 4.0 and other related policies,
based industrial intelligence across all manufacturing the approach seeks not only to be comprehensive
sectors, invest KRW8.4 trillion (USD6.91 billion) into but sustainable as well. The foresight to not only
the R&D of three key industries: non-memory chips, focus government investment on the development
future mobility, and biotechnology, as well as build of technologies but also on the talent and
2,000 AI-based factories and enact manufacturing infrastructure needed to support and sustain
innovation laws by 2030.298 development in the long-term is a clear strength.
Apart from channelling government budget, the In addition, the consideration and emphasis of
government’s emphasis on industry collaboration is ethics from the onset is commendable as it sets up
also a strong factor in its AI development. Samsung the policy space to discuss a fundamental concern
is leading the charge, first establishing its AI of society at large. For instance, the Mid to Long
research centre in 2017.299 Most recently, Samsung Term Master Plan in Preparation for the Intelligent
announced plans to invest USD22 billion into AI by Information Society is anchored on the concept of a
2020, and USD116 billion into non-memory chips to human-centred intelligent information society. This
power AI-enabled devices through to 2030.300 consideration of the human aspect of AI and other
intelligent technologies is also re-emphasised in the
Ethical AI is another major priority for the South 6th National Informatisation Master Plan (2018-2022).
Korean government, The Mid to Long Term Master Moving forward, the government also seeks to
Plan in Preparation for the Intelligent Information create a Charter of Ethics for developers and users
Society is anchored on the concept of a human- of intelligent technologies like AI, to reduce the
centred intelligent information society. This misuse of technology.303
consideration of the human aspect of AI and other
intelligent technologies is also re-emphasised in Apart from channelling government budget, the
the 6th National Informatisation Master Plan (2018- government’s emphasis on industry collaboration is
2022). Moving forward, the government also seeks also a strong factor in its AI development. In South
to create a Charter of Ethics for developers and Korea, the private sector plays an instrumental role
users of intelligent technologies like AI, to reduce in accelerating the AI industry. Samsung is leading
the misuse of technology.301 the charge, first establishing its AI research centre
in 2017.304 Most recently, Samsung announced plans
As competitive as South Korea’s strategy may be, to invest USD22 billion into AI by 2020, and USD116
it must also resolve long-standing institutional billion into non-memory chips to power AI-enabled
obstacles. It is, for instance, overcoming the devices through to 2030.305

Challenges
Lack of start-ups, domestic venture capital, and international collaboration
A large portion of the country’s AI developments have been concentrated in the activities of
industry incumbents and conglomerates, potentially hindering the innovation and growth of the AI
start-ups and the tech industry at large. While there are signs of improvement, with the investment
and engagement of smaller firms and start-ups—for example the AI Quantum Computing
Information Technology Research Centre set up by KAIST and funded by the government, domestic
telecommunication giant KT and three domestic SME tech firms—more can be done on this front to
engage smaller players in the market.306 This is compounded by the fact that the domestic venture
capital ecosystem is relatively underdeveloped in South Korea, resulting in a small number of
AI-focused start-ups to begin with.307 To alleviate this, the government could consider regulatory
changes that would encourage the development of venture capitalism in South Korea.
In addition, addressing the lack of experienced and competent AI talent will take time. Nurturing
domestic AI talent could also benefit from greater facilitation of international collaboration
and information exchange. Whilst many domestic universities and research institutes like the
Korea Advanced Institute of Science and Technology (KAIST), National Institute for Science and
Technology in Ulsan (UNIST), Pohang University of Science and Technology (POSTECH) and
Electronics and Telecommunications Research Institute (ETRI) have increasingly focused on AI and
machine learning, their scientific publications have had minimal exposure and influence outside
South Korea, and there is little international collaboration that promotes innovation and growth.308
This is one area that the government could work on in conjunction with the training of AI talents.

Thailand National AI Strategy No
Ministry of Information and
Communication Technology, Ministry
National Agency
of Digital Economy and Society
(MDES)
Key AI Organisations AI Technologies, OxygenAI, Jitta
E-commerce, Manufacturing,
Agriculture, Smart City, Vehicles,
Key Industry/Sector
Financial Services, Healthcare,
Education
Key AI Technology Machine Learning, Video Analytics,
Chatbots, Algorithms
Rank in EIU’s Automation Readiness
n/a
Index309
Rank in Oxford Insights’ Government
56 / 194
AI Readiness Index310
Six key AI principles are outlined in the draft:

AI must contribute to competitiveness and
sustainable development; it must comply
Ministry of Digital with law and international standards; it
Draft Ethical Guidelines Cross-
Economy and must be developed with accountability and
for AI (2019)311 sectoral
Society (MDES responsibility; it must uphold security and
privacy; it must promote equality, diversity,
inclusion, and fairness; and it must be made
reliable.
Office of the
The Twelfth National
National Economic Use advanced technologies such as AI, IoT,
Economic and Social Cross-
and Social smart devices and robotics to strengthen
Development Plan sectoral
Development existing production and services.
(2017-2021)312
Board (NESDB)
To develop Thailand into a smart and

Ministry of Digital
Cross- technology-driven economy. This policy also
Thailand 4.0 (2016)313 Economy and
sectoral encourages AI adoption in its Smart Cities
Society (MDES)
initiative.
Digital Economy
Digital Park Thailand Cross- Serve as the digital and innovation hub of the
Promotion Agency
(2018)314 sectoral ASEAN sub-region.
(DEPA)
Transform the Thai government into a digital

government and digitise all sectors including
Digital Government Digital Government
Public Sector agriculture, tourism, education, health and
Plan (2017-2021)315 Agency (DGA)
medicine, investment, disaster prevention,
and public administration.

Thailand has indeed seen tremendous growth in its
technology sector in recent years, and could very
soon lead Southeast Asia in terms of industrial AI
adoption. According to statistics from the Thailand
Board of Investment (BOI), shipments of industrial
robots increased by 133% to 7,500 units between
2013 and 2018, mainly driven by the automotive
and electronics and electrical appliance industries.
50% of manufacturers in Thailand are considering
adopting automation systems within the next three
years.322
At the same time, however, AI start-ups are relatively

new in Thailand and are still in the initial stage
of their AI journey.323 Dr Passakon Prathombutr
notes, “The main AI trend in Thailand is in the area
of natural language processing, which includes
Country Overview sentiment analysis and machine translation. There
is also the field of video content analysis, which
Thailand has yet to formulate an overarching focuses on facial recognition and license plate
national AI policy. For now, says Dr Passakon recognition (LPR).”324 In 2018, around 20 Thai
Prathombutr, Executive Vice President of the Digital enterprises – ranging from conglomerates to banks,
Economy Promotion Agency (DEPA), “Thailand is telecom operators and retailers – adopted some
focusing on capacity building, re-skilling, and up- type of AI solution, driven by a need to differentiate
skilling its manpower to prepare society for the wide their products and services.325 It is estimated that
impact of AI. This is an essential step to become a some 100 enterprises in Thailand will embrace AI in
regional and global hub for AI.”316 2019.326
In this context, most of Thailand’s AI initiatives Strengths

are embedded within policies and strategies
developed by different ministries and agencies to Innovation and development across multiple
bolster the development of AI-related technologies, sectors
including the Twelfth National Economic and Social
Development Plan (2017-2021),317 the Thailand 4.0 AI is being adopted across various industry verticals
policy,318 and Digital Park Thailand.319 in Thailand. The healthcare sector, already a major
economic asset for Thailand, is set to benefit from
Thailand 4.0, for example, emphasises the growth greater automation. The Thai Ministry of Public
and spread of advanced technologies such as AI, Health has started leveraging machine learning
IoT, smart devices, and robotics to transform existing and computer visualisation systems to help its
production and services into smart, value-added affiliated government agencies identify public health
production and services320 – thus diversifying the risks and disease hotspots and mitigate the risk of
economy and helping escape the “middle income epidemics.327 The Khon Kaen smart city initiative,
trap”. The Twelfth National Economic and Social meanwhile, is undertaking a Smart Health Project
Development Plan (2017-2021), meanwhile, is partly encompassing smart ambulances and preventive
centred around the use of advanced technologies healthcare services in a joint effort between local
such as AI, IoT, smart devices, and robotics to healthcare provides, academics, and the Thai
transform existing production and services into Government.328
smart, value-added production and services.321

AI is also used to predict babies’ health and with the Digital Economy Promotion Agency (DEPA)
development based on their parents’ profile to to create a learning curriculum that will produce
reduce medical costs for the government. To this 40,000 digital and high-tech workers by 2022.335
extent, the Ministry of Digital Economy and Society The DEPA is also working with Kasetsart University
(MDES) is setting up a “big data sandbox” – an to open the Digital Academy Thailand, aimed at
experimental regulatory ‘safe space’ that allows educating and developing digital skills in AI and
organisations to test digital products and services Data Science.
without putting data or systems at risk – for other
ministries to run similar experiments with their own DEPA also spearheaded the Digital Park Thailand,
datasets.329 which is expected to serve as the digital and
innovation hub of the ASEAN sub-region. A flagship
The banking sector plans to move to facial project of the Eastern Economic Corridor (ECC), it
recognition for electronic know-your-customer aims to create a favourable ecosystem to enable
regulations, as well as blockchain and machine the development of innovative solutions, for
learning for fraud detection. For Dr Passakon which USD160 million has been earmarked. An IoT
Prathombutr, “AI is shaping the future of the financial Institute will be located inside the park’s Digital
services industry in Thailand, with more local and Valley, that will use IoT and AI as key technologies
foreign Fintech companies implementing AI in for development.336 Likewise, the Thai People Map
their processes and operations.”330 The oil and gas and Analytic Platform (TPMAP) developed by the
industry is also implementing AI to detect dangerous National Electronics and Computer Technology
conditions for personnel, while retail businesses Centre (NECTEC) applies AI to reduce poverty in
use AI for customer experience and loyalty the country. NECTEC plans to initiate its own R&D
programmes.331 projects, such as National Language Processing to
enable the business sector to develop its own AI
Public-private partnerships to boost local talent usage model.337
and development
To further facilitate and enable AI use and

development, the Thai government has engaged
in public-private partnerships to improve the
skillset of its economy to meet its AI ambitions. For
instance, Thailand’s National Science, Technology
and Innovation (STI) Policy Office is working with
Google and the United Nations Economic and Social
Commission for Asia and the Pacific (UN-ESCAP) on
a joint AI for Social Good initiative, bringing together
researchers, policy makers, NGOs, and the private
sector from across the APAC region to discuss the
uses of AI for sustainable development.332
Likewise, the Ministry of Education is working with

the British Council and Chulalongkorn University’s
Computer Engineering Department to develop
an AI enabled system capable of evaluating
students’ speaking and writing abilities.333 The
Office Vocational Education Commission (OVEC) is
partnering with private sector entities to boost STEM
skills and spearhead the use of new technologies
like AI, automation and robotics.334 And Google,
Microsoft, Cisco, and Huawei are collaborating

Challenges
Lack of clear government guidelines
Although the Thai government aims to make better use of AI and prioritise the development
and adoption of AI, it has yet to develop clear strategies and guidelines at the national level.
Additionally, there is limited strategic coordination between the country’s existing national plans
to promote AI, as well as limited success metrics to evaluate and assess where gaps still exist
and need to be addressed.
Lack of AI-capable talent and R&D initiatives
Thailand currently lacks sufficient talent and R&D in AI technology. To address this gap, Thailand is
examining the possibility of adding AI to high school curricula to promote AI among the youth.338
Another challenge for Thailand is the fact that most workers are low- or mid-level employees,
meaning several of them may be replaced by automation, especially in the sectors of healthcare,
banking, and retail. As for high-level employees and senior management, new jobs can be found in
the field of data analysis, but preparing them to take on these new skillsets will take time.339

V.
Conclusion

Design Principles for Effective AI Policies
Policy-makers have an important role to play in

ensuring that AI becomes a tool for societal good
and does not simply exacerbate or aggravate
current socio-economic challenges in APAC.
From best practices to ethical guidelines, policy-
makers can help individuals, organisations, and
communities take advantage of the benefits of AI
and its associated innovations, while mitigating any
unintended or as-yet unidentified risks.
This examination of countries’ specific policy

approaches to AI suggests a number of design
principles that APAC policy-makers are looking into
when shaping ecosystems, preparing talent, and
building trust.
Design Principle 1:
Shaping AI Ecosystems
From investment and procurement schemes to

labour policies and open data mechanisms, the
public sector stands to play a significant role in
the expansion and adoption of AI technologies.
Governments that design sound, coherent national
plans around AI and allow their institutions to
execute these plans are one step ahead of
governments debating the nature or value of AI
systems.
Enable innovation by reducing red tape and First piloted in eight Chinese cities and regions, the
boosting entrepreneurship approach allowed the authorities to monitor and
evaluate the effectiveness of the reforms before
Across APAC, red tape remains a major obstacle implementing them on a wider scale.340 In case
for technology entrepreneurs. In most cases, of security or compliance concerns, regulatory
this is not due to governments being against the sandboxes – special regulatory provisions that
emergence of innovative businesses, but a natural allow businesses to test innovative technologies in
consequence of having outdated policies and/ a controlled environment while giving governments
or slow-moving bureaucracies. China recently room to make regulatory experiments – can be set
overcame this constraint by progressively testing up, as Singapore has done to encourage Fintech
and rolling out pro-innovation reform measures. innovation.341

Make data open and available is not hindered by the regulatory safeguards
designed to keep it secure.
AI technologies require a steady diet of reliable,
actionable, and secure data to learn and function. Strengthen multi-lateral/regional cooperation
But many industries across the region struggle to
track, capture, store, and aggregate operational APAC has copious AI ‘assets’, such as well-funded
data in a way that informs AI systems. In many centres of research, populous and data-generating
organisations, data sits in silos with fragmented consumer markets, and government expertise in
ownership. In others, vast quantities are collected coordinating value networks. Working together,
but never analysed. Governments can help governments, entrepreneurs, technology firms, and
organisations institute effective data governance research institutes operating in APAC can build a
protocols – including data anonymisation processes collaborative ecosystem that addresses common
– so that AI systems can better access them and development challenges.
learn from them. Governments can also bridge
data gaps by opening their own data; making Design clear, comprehensive government plans
government data freely available for everyone to
access, use, and republish without restrictions from AI, like all other digital technologies able to
copyright, patents, or other mechanisms of control transform entire economic sectors, is most
can help AI technologies advance in the interest of effective when guided by sound, coherent policies.
citizens and consumers alike. In this regard, South The digital age is the domain of innovative tech
Korea is the top performing country in the APAC companies but driving the adoption of disruptive
region. Both the Open Data Barometer342 and the products and services requires the support of
OECD343 note its willingness to make government governments with both vision and capability.
data available on national portals as well as useful to Overall, it seems that public spending is poised to
the delivery of government services and benefits. become as important as governments’ planning and
implementation mechanisms. In Australia, a number
Facilitate cross-border flows and regional of AI professionals have called for more public
cooperation spending in AI.346
Cross-border data flows are imperative to drive Address inequalities that prevent wider AI
AI growth.344 In the age of hyper-connected adoption/deployment
people, devices, and platforms, data flows across
multiple locations and jurisdictions are very difficult The Salesforce AI Readiness Index shows marked
to contain within physical and virtual borders. differences between mature and emerging
Facilitating the movement of data requires effective economies, suggesting that the digital divide
and consistent cooperation between governments that has long characterised the region can and
that share a common goal. Initiatives such as the may already be carrying over to the AI sector.347
APEC Cross-Border Privacy Rules (CBPR) system According to the World bank, the prolongation of
can ensure cross-border flows of data and personal the digital divide is to be expected, as the spread of
information are as useful to AI technologies as they digital technologies has yet to equate to the spread
are secure for users. In this regard, Japan stands of digital dividends.348 For the ITU, this divide is
out as both a regional and global leader. According more of a temporary phenomenon that will naturally
to the Salesforce Cross-Border Data Flows Index taper off as AI technologies gain in maturity and are
(CBDFI), Japan has the least restrictive and the most used for social good.349
consistent approach to cross-border data flows.345
Japan has indeed ramped up efforts to create a
regulatory environment that both enables and
protects the free flow of data across boundaries –
ensuring that the movement of data across borders

Design Principle 2: Preparing AI Talent
AI technologies require advanced skills and workers
who can develop and maintain complex AI systems
and applications. But recent estimates find that there
are only 300,000 AI engineers, researchers, and
practitioners worldwide, when several millions of
them will be needed over the next two decades.350
Across the board, companies report that finding the
right talent is the biggest hurdle they face in trying
to integrate AI into their existing operations. And the
shortage is even more acute in the APAC region.351
Adapt educational institutions and what they Support upskilling and lifelong learning schemes
teach
According to some estimates, AI will push as many
Curriculum reform of both compulsory and post- as 375 million workers globally – or 14% of the
compulsory education is gaining traction in most of global workforce – to reskill or make a transition to
the countries in the Index. In Singapore, coding has new occupational categories.354 Lifelong learning
been introduced to primary and secondary school and upskilling schemes are already helping workers
curricula, as well as STEM coursework in secondary upgrade and diversify their skills. Such programmes
and vocational training programmes.352 Such could be expanded to include AI-specific skills
initiatives could be used to introduce students to such as applied statistics, computational thinking,
the more abstract facets of AI, including the role of graphical modelling, robotics, programming
ethics in computer science, the influence of biases languages, and cognitive science theory.
in society and in databases, and the importance of
data quality when using data-driven technologies. Prepare workers to the challenges of tomorrow
Use AI to make education better A recent study shows that some 60% of occupations
have at least 30% of constituent work activities
Big data and analytics can be used to improve that could be automated, suggesting that many
both learning and teaching experiences. Capturing workers will be impacted by the rise of automation
data on student demographics, school attributes, technologies.355 But even with automation, the
and individual trajectories can help decision- demand for skilled and specialised workers will
makers optimise the way they allocate resources or keep on growing, as productivity gains move the
adjust policies. AI-based Intelligent Tutor Systems labour market towards a new class of “smart”,
(ITS), for example, can deliver precise support to knowledge-intensive jobs. To meet the needs,
students. Universities in Singapore and Malaysia are expectations, and demands of a fast-evolving labour
experimenting with predictive software that detects market, workers will have to learn to work with
the likelihood of students dropping out of a class.353 machines instead of against them.
AI systems can also help teachers improve their
workflows, by relieving them of some of the more
routine, time-consuming tasks of teaching, including
grading and record-keeping.

Design Principle 3: Building Trust in AI
Privacy and security are central to the development
and deployment of AI technologies. As AI grows
increasingly complex and widespread, it will become
inescapable for AI-driven organisations to comply
with laws that regulate the way individuals’ and
organisations’ data is collected, stored, used, and
protected. In this context, trust will be crucial for
the rise of AI in day-to-day functions. Not only must
algorithms be understandable, transparent, reliable,
and trustworthy, the organisation using it must be
accountable in case it leads to errors or misuses.
Enforce privacy principles and practices systems to be transparent; not only how, but also
why a decision or an action was reached. Making
“Privacy by Design” comprises a number of AI explainable in this manner allows consumers to
technical and organisational measures at each better understand how AI systems work and affect
stage of the data collection and processing chain. them.
It pushes organisations to consider the privacy
and security requirements of their data-dependent Control the way data is collected and used
products, systems, and operations every step of
the way. It also includes security measures such as There are many ways to ensure personal data
access controls, audit logs, encryption, and data is used in the way it is meant to be used. Data
segregation to keep personal data separate from anonymisation – deleting or encrypting personally
other forms of data. Conducting regular privacy identifiable information to make the data irreversibly
impact assessments (PIA) can also help identify and untraceable and unreproducible – allows AI systems
mitigate privacy risks before the actual processing to use data without using or divulging any personal
of personal data. While AI systems may involve or private information, mitigating the risks of data
innovative, complex, and sometimes unexpected/ breaches and accidental disclosures. Data limitation
unintended uses of personal data, the use of PIAs policies, meanwhile, force organisations to limit
help organisations better assess the risks and the use of personal data for the purpose for which
impacts involved in the processing of personal data. it was originally collected, and for which consent
was provided. Finally, data minimisation reduces
Make transparency a key feature of AI products the amount of data collected and processed by
and services establishing at the onset what data is relevant and
necessary for a given purpose.
Increasingly complex digital products and services
require increasingly complex uses of data. Recent Enable ethical, unbiased AI
scandals involving the lack of consent to use
personal data for decision-making have driven For AI to grow and deliver on its promises, it must
individuals to demand more transparency on what be designed and deployed in a manner that earns
personal data is processed and who the data and keeps the trust of individuals, organisations,
is given to.356 As the number of self-managing, and governments alike. Making AI trustworthy
hyper-connected devices grows across homes, entails embedding ethical principles into AI systems,
offices, and even public spaces, it is crucial for AI ensuring they act in a fair, inclusive, and humane

manner.357 One way of ensuring AI technologies
positively impact the world around them is to
identify and minimise the risk of bias, especially
when AI systems are in a position to influence
recommendations and predictions.
AI algorithms are only as good as the data they
are fed to learn, which means it is up to humans to
manually take fairness and diversity into account as
datasets are built.
Use AI to strengthen cybersecurity
Researchers in the United States and China have

successfully used commercially available AI systems
to dial phones and open websites without the
knowledge of the AI systems’ users, a short step
to more malicious intents, such as unlocking doors
and withdrawing funds.358 Compromised AI systems
are a real risk, especially when the absence of
human intervention can allow it to go undetected.
But AI can be used to patch its own weaknesses;
AI can improve existing detection and response
capabilities, as well as create new preventative
defence protocols. Companies can use AI platforms
to streamline complex, manual, or time-consuming
inspection processes, allowing specialists to devote
their attention to more immediate threats.
Use AI for good and teach AI to do good
Collaborating with private-sector organisations,

governments can share AI tools and resources,
datasets, and supporting knowledge and expertise
to address pressing sustainability challenges and
help vulnerable populations. Governments can
also drive and support efforts aimed at better
understanding and addressing the legal and ethical
issues that AI technologies create. The recent
work of the IIC with the UN-ESCAP – Artificial
Intelligence for Economic Growth and Social Good
– demonstrates the importance of looking at the
way in which policy makers might understand and
address the issues raised by the increasing use of AI
as both an economic and societal enabler.359

VI.
References

1. CIO, www.cio.com/article/3311756/how-is-artificial-intelligence-benefiting-industries-throughout-southeast-asia.html
2. Insurance Business Asia, www.insurancebusinessmag.com/asia/news/breaking-news/ai-to-reorganise-asias-insurance-landscape-170431.aspx
3. Tech in Asia, www.techinasia.com/alibaba-ai-et-brain-agriculture
4. PwC, www.pwc.com/gx/en/issues/data-and-analytics/publications/artificial-intelligence-study.html
5. This encompasses both internal R&D efforts and a steady stream of acquisitions. McKinsey, www.mckinsey.com/~/media/McKinsey/Featured%20Insights/Artificial%20Intelligence/AI%20
and%20SE%20ASIA%20future/Artificial-intelligence-and-Southeast-Asias-future.ashx
6. UNDP, www.undp.org/content/undp/en/home/blog/2019/Using_AI_to_help_achieve_Sustainable_Development_Goals.html
7. SCMP, www.scmp.com/tech/science-research/article/2145568/can-trumps-ai-summit-match-chinas-ambitious-strategic-plan
8. Asian Scientist, www.asianscientist.com/2018/03/topnews/southeast-asia-ai-research-clarivate
9. Open Gov Asia, www.opengovasia.com/south-korea-to-invest-2-2-trillion-won-in-bid-to-seize-the-lead-in-ai-technology-by-2022
10. Tech in Asia, www.techinasia.com/indonesia-ai-digital-disruption
11. Enterprise IT News, www.enterpriseitnews.com.my/artificial-intelligence-in-malaysia-challenged
12. Thailand Board of Investment, www.prnewswire.com/news-releases/thailand-promotes-ai-robotics-technology-to-spur-industry-4-0-readiness-300854969.hml
13. University of Santo Tomas, www.ust.edu.ph/academics/brain-computer-typing-interface-bags-top-bpi-dost-project-of-the-year-award-for-nieles-of-ust-engineering
14. The Tech Panda, https://thetechpanda.com/drishti-accenture/23373
15. UN-ESCAP, www.unescap.org/resources/updated-analysis-broadband-infrastructure-asia-pacific
16. International Telecommunication Union (ITU), www.itu.int/pub/S-POL-BROADBAND.20-2019
17. AT Kearney, www.southeast-asia.atkearney.com/documents/10192/6986374/ASEAN+Digital+Revolution.pdf/86c51659-c7fb-4bc5-b6e1-22be3d801ad2
18. Bain & Company, www.bain.com/insights/advancing-towards-asean-digital-integration
19. Deloitte, www2.deloitte.com/us/en/insights/focus/cognitive-technologies/ai-investment-by-country.html
20. McKinsey, www.mckinsey.com/featured-insights/future-of-work/ai-automation-and-the-future-of-work-ten-things-to-solve-for
21. Salesforce, www.salesforce.com/ap/form/conf/ai-readiness-index
22. OECD, www.oecd.org/going-digital/ai/principles
23. OECD, www.oecd.org/going-digital/ai/about-the-oecd-ai-policy-observatory.pdf
24. G20, www.mofa.go.jp/mofaj/files/000486596.pdf

25. International Telecommunication Union (ITU), www.itu.int/en/ITU-T/AI/2018/Pages/default.aspx
26. Channel news Asia, www.channelnewsasia.com/news/singapore/nea-alphabet-inc-s-verily-team-up-to-fight-dengue-with-ai-10731556
27. Asian Scientist, www.asianscientist.com/2017/11/pharma/ai-microscope-motic
28. Hindustan Times, www.hindustantimes.com/tech/google-launches-ai-project-in-thailand-to-screen-for-diabetic-eye-disease/story-d6rKjfTDxeANwwasfj6ZZP.html
29. Nhan Dan, https://en.nhandan.com.vn/scitech/sci-tech/item/7853402-vietnam%E2%80%99s-first-ai-learning-system-launched.html
30. SCMP, www.scmp.com/week-asia/business/article/2118345/artificial-intelligence-rise-southeast-asia-helping-everyone
31. Jakarta Post, www.thejakartapost.com/news/2019/03/28/tokopedia-ui-launch-ai-research-center.html
32. The Tech Panda, https://thetechpanda.com/drishti-accenture/23373
33. MIMOS, www.mimos.my/huawei-mimos-to-identify-safety-and-smart-city-solutions
34. PwC, www.pwc.com/gx/en/sustainability/assets/ai-for-the-earth-jan-2018.pdf
35. ASEAN up, https://aseanup.com/southeast-asia-digital-social-mobile
37. UN-ESCAP, www.unescap.org/resources/updated-analysis-broadband-infrastructure-asia-pacific
38. AT Kearney, www.southeast-asia.atkearney.com/documents/10192/6986374/ASEAN+Digital+Revolution.pdf/86c51659-c7fb-4bc5-b6e1-22be3d801ad2
40. internet Society, www.internetsociety.org/wp-content/uploads/2017/11/APAC-Issue-paper-Mobile-internet.pdf
41. SCMP, www.scmp.com/comment/insight-opinion/article/2130501/aseans-digital-divide-obstacle-singapores-regional-market
42. GSMA, www.gsma.com/r/mobileeconomy/asiapacific
43. World Bank, www.worldbank.org/en/topic/poverty/overview
44. The Conversation, https://theconversation.com/digital-innovation-in-indonesia-inequalities-between-the-east-and-the-west-95218
45. National Geographic, www.nationalgeographic.com/science/2018/07/news-internet-underwater-sea-level-rise
46. The Diplomat, https://thediplomat.com/2016/09/closing-asias-digital-divide

47. Oxford Economics, https://3er1viui9wo30pkxh1v2nh4w-wpengine.netdna-ssl.com/wp-content/uploads/prod/sites/43/2018/07/Harnessing_the_Digital_Dividend_SEA_8cases_20180724.pdf
48. Microsoft, https://blogs.microsoft.com/on-the-issues/2010/11/11/economics-of-the-cloud
49. Ibid.
50. Asia Cloud Computing Association (ACCA), www.asiacloudcomputing.org/research/2018-research/cri2018
51. Business World, www.bworldonline.com/acca-pushes-improvements-to-countrys-cloud-first-policy
52. Japan Times, www.japantimes.co.jp/news/2019/06/28/national/abe-heralds-launch-osaka-track-framework-free-cross-border-data-flow-g20
53. Discover Magazine, http://discovermagazine.com/2018/dec/outsmarting-outbreaks
54. University of Washington, https://jsis.washington.edu/news/chinese-data-localization-law-comprehensive-ambiguous
55. Baker McKenzie, www.bakermckenzie.com/-/media/files/insight/publications/2018/07/al_generaldatalocalizationrequirements_july2018.pdf?la=en
56. Vantage Asia, www.vantageasia.com/new-regulation-ott-activities-indonesia
57. International Association of Privacy Professionals (IAPP), https://iapp.org/news/a/gdpr-matchup-south-koreas-personal-information-protection-act
58. Information Technology & Innovation Foundation (ITIF), https://itif.org/publications/2017/05/01/cross-border-data-flows-where-are-barriers-and-what-do-they-cost
59. National Institute of Standards and Technology (NIST), https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-60v1r1.pdf
60. Attorney-General’s Department, www.protectivesecurity.gov.au/information/sensitive-classified-information/Documents/pspf-infosec-08-sensitive-classified-information.pdf
61. Ibid.
62. McKinsey, www.mckinsey.com/~/media/mckinsey/featured%20insights/Future%20of%20Organizations/What%20the%20future%20of%20work%20will%20mean%20for%20jobs%20skills%20
and%20wages/MGI-Jobs-Lost-Jobs-Gained-Report-December-6-2017.ashx

63. AlphaBeta, www.alphabeta.com/our-research/the-automation-advantage
64. Future of Life Institute, https://futureoflife.org/ai-open-letter
65. EIU, www.automationreadiness.eiu.com
66. ZDNet, www.zdnet.com/article/singapore-aims-to-build-up-ai-skills-for-digital-economy/
67. Straits Times, www.straitstimes.com/tech/coding-to-be-made-compulsory-for-all-upper-primary-pupils-next-year
68. MSIT, www.msit.go.kr/web/msipContents/contentsView.do?cateId=mssw311&artId=1454538
69. Malay Mail, www.malaymail.com/news/malaysia/2019/05/19/report-year-four-students-to-learn-ai-robotics-and-computer-programming-in/1754364
70. Open Gov Asia, www.opengovasia.com/malaysias-ministry-of-education-to-introduce-ai-robotics-and-computer-programming-in-school
71. EIU, https://automationreadiness.eiu.com/whitepaper
72. Forbes, www.forbes.com/sites/bernardmarr/2018/06/25/the-ai-skills-crisis-and-how-to-close-the-gap/#1174db3931f3
73. e27, https://e27.co/much-ai-talent-asia-pacific-look-data-trends-20171101
74. The Verge, www.theverge.com/2017/12/5/16737224/global-ai-talent-shortfall-tencent-report
75. SCMP, www.scmp.com/tech/china-tech/article/2123949/worlds-most-populous-country-lacks-one-thing-it-needs-become-ai
76. Ibid.
77. A*Star Singapore, www.a-star.edu.sg/News-and-Events/A-STAR-INNOVATE/Index/ID/7917
78. Tech Wire Asia, https://techwireasia.com/2019/08/with-salaries-for-ai-talent-soaring-daikin-invests-in-upskilling-staff
79. Email interview conducted by TRPC, August 2019.
80. Medium, https://medium.com/syncedreview/chinas-ai-schools-are-accepting-applications-here-s-a-list-5a568e1e31a1
81. NDTV, www.ndtv.com/education/artificial-intelligence-study-options-in-india-and-abroad-1717859
82. Microsoft, https://news.microsoft.com/apac/2019/09/26/grabbing-opportunities-for-growth-through-reskilling-in-asia-pacific
83. Use AI in a way that respects human rights, and share the benefits brought by it www.salesforce.com/ap/form/conf/ai-readiness-index
84. Medium, https://medium.com/airbnb-engineering/how-airbnb-uses-machine-learning-to-detect-host-preferences-18ce07150fa3
85. OCBC, www.ocbc.com/group/media/release/2018/ocbc-launches-ai-lab.html
86. SCMP, www.scmp.com/news/china/politics/article/2170834/chinas-sharp-eyes-surveillance-system-puts-security-focus-public
87. Malay Mail, www.malaymail.com/news/malaysia/2019/01/02/penang-launches-countrys-first-facial-recognition-cctv-surveillance/1708422

88. Reuters, www.reuters.com/article/us-singapore-surveillance/singapore-to-test-facial-recognition-on-lampposts-stoking-privacy-fears-idUSKBN1HK0RV
89. The Star, www.thestar.com.my/tech/tech-news/2019/02/26/chinese-police-test-gaitrecognition-technology-from-ai-startup-watrix-that-identifies-people-based-o
90. Government Gazette, www.ratchakitcha.soc.go.th/DATA/PDF/2562/A/069/T_0052.PDF
91. Strategy & Business, www.strategy-business.com/article/The-Future-of-Artificial-Intelligence-Depends-on-Trust?gko=af118
92. CMS Law, https://cms.law/en/GBR/Publication/The-future-landscape-for-AI-claims-around-the-world
93. METI, www.meti.go.jp/english/press/2018/0615_002.html
94. BAAI, www.baai.ac.cn/blog/beijing-ai-principles
95. CSIRO, https://consult.industry.gov.au/strategic-policy/artificial-intelligence-ethics-framework
96. Ministry of Transport, www.mot.gov.sg/news-centre/news/Detail/Committee-on-Autonomous-Road-Transport-for-Singapore
97. www.mas.gov.sg/~/media/MAS/News%20and%20Publications/Monographs%20and%20Information%20Papers/FEAT%20Principles%20Final.pdf
98. IMDA, www2.imda.gov.sg/news-and-events/Media-Room/Media-Releases/2019/inaugural-meeting-of-the-advisory-council-on-the-ethical-use-of-artificial-intelligence-and-data
99. IDC, www.idc.com/getdoc.jsp?containerId=prUS44291818
100. Ibid.
101. Pitchbook, https://pitchbook.com/news/reports/1q-2018-pitchbook-analyst-note-real-potential-for-ai
102. This encompasses both internal R&D efforts and a steady stream of acquisitions. McKinsey, www.mckinsey.com/~/media/McKinsey/Featured%20Insights/Artificial%20Intelligence/AI%20
and%20SE%20ASIA%20future/Artificial-intelligence-and-Southeast-Asias-future.ashx
103. Ibid.
104. CIFAR, www.cifar.ca/cifarnews/2018/12/06/building-an-ai-world-report-on-national-and-regional-ai-strategies
105. Towards Data Science, https://towardsdatascience.com/4-factors-influencing-chinas-ai-dominance-3e432ea55091
106. SCMP, www.scmp.com/tech/article/2126755/china-building-giant-21-billion-research-park-dedicated-developing-artificial
107. Open Gov Asia, www.opengovasia.com/chinas-ministry-of-education-unveils-five-year-ai-training-program-for-universities
108. Allen Institute for Artificial Intelligence (AI2), https://medium.com/ai2-blog/china-to-overtake-us-in-ai-research-8b6b1fe30595

109. Massachusetts Institute of Technology (MIT), http://news.mit.edu/2019/first-ai-policy-congress-0118
110. Ibid.
111. National Research Foundation, www.nrf.gov.sg/programmes/artificial-intelligence-r-d-programme
112. Economic Development Board (EDB), www.edb.gov.sg/en/how-we-help/incentives-and-schemes.html
113. Monetary Authority of Singapore (MAS), www.mas.gov.sg/publications/monographs-or-information-paper/2018/FEAT
114. MIT Technology Review, www.technologyreview.com/s/612654/asias-ai-agenda-the-ecosystem
115. Civil Service College (CSC), www.csc.gov.sg/articles/editorial-issue-21
116. MIT Technology Review, www.technologyreview.com/s/612654/asias-ai-agenda-the-ecosystem
117. Ibid.
118. All laws and regulations examined in this section are accurate as of November 2019.
119. European Commission, https://ec.europa.eu/digital-single-market/en/news/eu-member-states-sign-cooperate-artificial-intelligence
120. European Commission, https://ec.europa.eu/digital-single-market/en/news/communication-artificial-intelligence-europe
121. European Commission, https://ec.europa.eu/futurium/en/ai-alliance-consultation/guidelines#Top
122. European Commission, https://ec.europa.eu/digital-single-market/en/news/policy-and-investment-recommendations-trustworthy-artificial-intelligence
123. G20, www.mofa.go.jp/files/000486596.pdf
124. Government of Canada, www.canada.ca/en/innovation-science-economic-development/news/2019/05/declaration-of-the-international-panel-on-artificial-intelligence.html
125. OECD, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449

126. OECD, www.oecd.org/going-digital/ai/principles
127. OECD, www.oecd.org/going-digital/ai/about-the-oecd-ai-policy-observatory.pdf
128. Ministry of Internal Affairs and Communications (MIC), http://events.science-japon.org/dlai17/doc/MIC%20-%20France-Japan%20Symposium%2020171025.pdf
129. New Energy and Industrial Technology Development Organisation (NEDO), www.nedo.go.jp/content/100865202.pdf
130. Japan Government, www.japan.go.jp/abenomics/_userdata/abenomics/pdf/society_5.0.pdf
131. Ministry of Internal Affairs and Communications (MIC), www.soumu.go.jp/main_content/000507517.pdf
132. Ministry of Economy, Trade, and Industry (METI), www.meti.go.jp/english/report/data/wp2017/wp2017.html
133. HRM Asia, http://hrmasia.com/one-of-worlds-most-advanced-economies-needs-to-train-up-ai-talent-fast
134. White House, www.whitehouse.gov/presidential-actions/executive-order-maintaining-american-leadership-artificial-intelligence
135. White House, www.whitehouse.gov/articles/accelerating-americas-leadership-in-artificial-intelligence
136. Straits Times, www.straitstimes.com/singapore/singapore-releases-framework-on-ethical-use-of-ai
137. Search Enterprise AI, https://searchenterpriseai.techtarget.com/feature/AI-regulation-stirs-as-unrestricted-AI-booms-in-China
138. The Edge Markets, www.theedgemarkets.com/article/alibaba-and-malaysia-reaffirm-partnership
139. Forbes, www.forbes.com/sites/cognitiveworld/2018/09/07/is-south-korea-poised-to-be-a-leader-in-ai/#5479cb59fa2f
140. Business Australia, www.business.gov.au/assistance/cooperative-research-centres-programme
141. Konrad-Adenauer-Stiftung (KAS), www.kas.de/documents/288143/4843367/panorama_digital_asia_v5a_Mitra.pdf/2309433a-f1dc-5c59-18be-20375225ea67
142. OECD, www.oecd.org/southeast-asia/events/regional-forum/Forum_Note_Digital_Transformation_STI.pdf
143. Open Gov Asia, www.opengovasia.com/indonesia-drafts-personal-data-protection-act
144. New Straits Times, www.nst.com.my/news/nation/2017/11/300540/najib-wants-value-innovative-centre-established-ministries-departments
145. IMDA, www.gov.sg/~/sgpcmedia/media_releases/imda/press_release/P-20180605-1/attachment/Artificial%20Intelligence%20Governance%20and%20Ethics%20Initiatives.pdf
146. Korea Herald, www.koreaherald.com/view.php?ud=20180620000688
147. Government News, www.governmentnews.com.au/government-should-lead-ai-certification-finkel
148. Deloitte, www2.deloitte.com/au/en/pages/technology/articles/how-countries-are-pursuing-ai-advantage.html
149. Financial Review, www.afr.com/technology/fears-australia-will-be-left-behind-with-minimal-ai-spending-20190116-h1a4oo

151. Oxford Insights, www.oxfordinsights.com/ai-readiness2019
152. On-the-ground intelligence gathered by TRPC, October 2019.
153. Department of Industry, www.industry.gov.au/sites/g/files/net3906/f/July%202018/document/pdf/national-innovation-and-science-agenda-report.pdf
154. Department of Industry, www.industry.gov.au/sites/default/files/2018-12/australias-tech-future.pdf
155. Office of the National Data Commissioner, www.datacommissioner.gov.au/advisory-council
156. Australian Government, www.business.gov.au/-/media/Business/CRC/Cooperative-research-centres-projects-factsheet-artificial-intelligence-PDF.
pdf?la=en&hash=087AD41AFC84DDA6AA8996C60E38C03F38491A67
157. Computer World, www.computerworld.com.au/article/640926/budget-2018-government-seeks-boost-australian-ai-capabilities
161. BCG, www.linkedin.com/pulse/where-australias-national-ai-strategy-miguel-carrasco
162. Australian Institute for Machine Learning (AIML), www.adelaide.edu.au/aiml/system/files/media/documents/2019-05/AIML%20final%20release.pdf
163. ZDNet, www.zdnet.com/article/commercialisation-opportunity-for-data61s-data-airlock-system-as-interest-ramps-up
164. Queensland Government, https://advance.qld.gov.au/ai-hub
165. Australian Institute for Machine Learning (AIML), www.adelaide.edu.au/aiml/artificial-intelligence
166. Melbourne has put forward AUD1 million (USD700,150) for the National Centre for AI. www.opengovasia.com/melbourne-to-welcome-national-artificial-intelligence-centre
167. ZDNet, www.zdnet.com/article/university-of-sydney-researchers-to-use-ai-to-help-prevent-heart-attacks/
168. ZDNet, www.zdnet.com/article/microsoft-backs-six-aussie-environment-focused-ai-projects/
169. Open Gov Asia, www.opengovasia.com/federal-court-turns-to-ai-to-predict-asset-split-after-relationship-breakdown/
170. ZDNet, www.zdnet.com/article/nab-doubles-down-on-value-of-data-and-analytics-with-new-in-house-guilds/
171. AHRC, https://tech.humanrights.gov.au/sites/default/files/2019-02/AHRC_WEF_AI_WhitePaper2019.pdf

172. Consultation Hub, https://consult.industry.gov.au/strategic-policy/artificial-intelligence-ethics-framework
173. Standards Australia, www.standards.org.au/getmedia/aeaa5d9e-8911-4536-8c36-76733a3950d1/Artificial-Intelligence-Discussion-Paper-(004).pdf.aspx
174. ACOLA, https://acola.org/wp-content/uploads/2019/07/hs4_artificial-intelligence-report.pdf
175. HMI, https://hmi.anu.edu.au
176. Salesforce, www.salesforce.com/au/blog/2019/09/why-australia-needs-an-ethical-ai-framework.html
177. Salesforce, www.salesforce.com/au/form/pdf/accelerating-australias-digital-transformation
179. Salesforce, www.salesforce.com/ap/form/conf/ai-readiness-index
180. ZDNet, www.zdnet.com/article/china-identifies-17-key-areas-to-make-ai-breakthroughs
183. State Council, https://flia.org/wp-content/uploads/2017/07/A-New-Generation-of-Artificial-Intelligence-Development-Plan-1.pdf
184. China Briefing, www.china-briefing.com/news/made-in-china-2025-explained
185. China Daily, www.chinadaily.com.cn/a/201709/18/WS59c0aa57a310a84d86d1565b.html
186. Open Gov Asia, www.opengovasia.com/chinese-government-sets-out-specific-targets-in-three-year-action-plan-on-artificial-intelligence
187. China Money Network, www.chinamoneynetwork.com/2018/09/19/china-money-network-announces-china-ai-top-50-ranking-at-summer-davos-in-tianjin
188. China Banking News, www.chinabankingnews.com/2018/09/20/chinas-top-50-artificial-intelligence-companies-revealed

189. CB Insights, www.cbinsights.com/research/china-artificial-intelligence-investment-startups-tech/
190. SCMP, www.scmp.com/comment/insight-opinion/article/2142641/chinas-ai-dream-well-its-way-becoming-reality
191. McKinsey, www.mckinsey.com/featured-insights/artificial-intelligence/kai-fu-lees-perspectives-on-two-global-leaders-in-artificial-intelligence-china-and-the-united-states
192. SPPM, www.sppm.tsinghua.edu.cn/eWebEditor/UploadFile/China_AI_development_report_2018.pdf
193. AI Index, http://cdn.aiindex.org/2018/AI%20Index%202018%20Annual%20Report.pdf
194. Asia Times, www.asiatimes.com/2019/02/article/shanghai-eyes-long-term-ai-development-fund/
195. SCMP, www.scmp.com/tech/innovation/article/2146428/tianjin-city-china-eyes-us16-billion-fund-ai-work-dwarfing-eus-plan
196. Intel, https://newsroom.intel.com/articles/intel-baidu-continue-collaboration-across-ai-ad-5g/#gs.wp39ov
197. Medium, https://medium.com/syncedreview/tencents-new-medical-ai-lab-targets-parkinson-s-715c5a1b68f2
198. China Daily, www.chinadaily.com.cn/a/201905/10/WS5cd4f338a3104842260baf95.html
199. BAAI, https://baip.baai.ac.cn/en?fbclid=IwAR2HtIRKJxxy9Q1Y953H-2pMHl_bIr8pcsIxho93BtZY-FPH39vV9v9B2eY
200. SPPM, www.sppm.tsinghua.edu.cn/eWebEditor/UploadFile/China_AI_development_report_2018.pdf

201. Nikkei Asian Review, https://asia.nikkei.com/Business/Business-trends/Chinese-tech-salaries-jump-in-global-race-for-talent

205. Makarim, www.makarim.com/uploads/230978_M&T%20Advisory%20-%20New%20Govt%20Reg%20on%20Organization%20of%20Electronic%20Systems%20and%20Transactions%20
(October%202019).pdf
206. AT Kearney, www.kemenperin.go.id/download/18280/Making-Indonesia-4.0---Bahan-AT-Kearney-Sosialisasi-tentang-Industry-4.0-(update)

207. Chandler Nguyen, www.chandlernguyen.com/blog/2019/02/05/9-key-facts-about-china-digital-landscape
208. Email interview conducted by TRPC, October 2019.

209. Jakarta Post, www.thejakartapost.com/news/2018/09/05/indonesia-leads-asia-pacific-in-ai-implementation-study-shows.html
210. IDC, www.idc.com/getdoc.jsp?containerId=AP43418218
212. McKinsey, www.mckinsey.com/~/media/McKinsey/Featured%20Insights/Artificial%20Intelligence/AI%20and%20SE%20ASIA%20future/Artificial-intelligence-and-Southeast-Asias-future.ashx

213. Email interview conducted by TRPC, October 2019.
214. Chandler Nguyen, www.chandlernguyen.com/blog/2019/02/05/9-key-facts-about-china-digital-landscape
215. Appier, www.appier.com/news/appiers-new-survey-on-ai-adoption-in-asia-pacific-indonesia-leads-the-pack-in-ai-implementation
216. Jakarta Post, www.thejakartapost.com/news/2019/02/04/bukalapak-launches-innovation-center-at-itb.html
217. Jakarta Post, www.thejakartapost.com/news/2019/03/28/tokopedia-ui-launch-ai-research-center.html
218. IDC, www.idc.com/getdoc.jsp?containerId=prAP43564418
219. McKinsey, www.mckinsey.com/~/media/McKinsey/Featured%20Insights/Artificial%20Intelligence/AI%20and%20SE%20ASIA%20future/Artificial-intelligence-and-Southeast-Asias-future.ashx
220. Speed Test, www.speedtest.net/global-index

222. Jakarta Globe, https://jakartaglobe.id/context/jokowi-wants-a-new-generation-of-top-talents-to-take-indonesia-to-developed-country-status

225. IDC, www.idc.com/getdoc.jsp?containerId=prAP45599119

226. NST, www.nst.com.my/news/nation/2019/04/475361/mdec-complete-national-ai-framework-year-end
227. Ministry of International Trade and Industry (MITI), www.miti.gov.my/miti/resources/National%20Policy%20on%20Industry%204.0/Industry4WRD_Final.pdf
228. MDEC, https://mdec.my/about-malaysia/government-policies/national-bda-framework/

229. NST, www.nst.com.my/news/nation/2019/04/475361/mdec-complete-national-ai-framework-year-end
230. The Star, www.thestar.com.my/business/business-news/2019/07/05/mdec-to-set-up-ai-unit
232. Asia IoT, http://iotbusiness-platform.com/blog/intelligent-transport-for-malaysias-smart-city-vision
233. My DFTZ, https://mydftz.com
234. MIMOS, www.mimos.my/huawei-mimos-to-identify-safety-and-smart-city-solutions

236. Analytics India Mag, www.analyticsindiamag.com/a-quick-look-at-5-deep-tech-startups-in-malaysia-with-big-ambitions/
239. Digital News Asia, www.digitalnewsasia.com/digital-economy/adoption-artificial-intelligence-rise-asean
241. Malay Mail, www.malaymail.com/news/malaysia/2019/08/08/personal-healthcare-now-made-easy-with-prudentials-pulse-app/1778920
242. Open Gov Asia, www.opengovasia.com/malaysia-must-push-for-greater-ai-adoption/
243. Alibaba Cloud, www.alibabacloud.com/press-room/alibaba-cloud-launches-malaysia-city-brain-to-enhance-city-management
244. Xin Hua Net, www.xinhuanet.com/english/2019-07/31/c_138272650.htm
245. Computer Weekly, www.computerweekly.com/news/252462380/Malaysia-to-get-its-first-AI-park
246. Analytics India Mag, www.analyticsindiamag.com/a-quick-look-at-5-deep-tech-startups-in-malaysia-with-big-ambitions/
247. Avanade, www.avanade.com/ms-my/solutions/analytics-and-ai/artificial-intelligence
248. Crayon, www.crayon.com/en-MY/AI-Centre-of-Excellence/

251. Smart Nation Singapore, www-smartnation-sg-admin.cwp.sg/docs/default-source/default-document-library/national-ai-strategy-summary.pdf

252. Business Times, www.businesstimes.com.sg/banking-finance/sff-x-switch-2019/mas-co-creates-framework-with-financial-sector-for-responsible-use
253. PDPC, www.pdpc.gov.sg/Resources/Model-AI-Gov
254. MAS, www.mas.gov.sg/~/media/MAS/News%20and%20Publications/Monographs%20and%20Information%20Papers/FEAT%20Principles%20Final.pdf
255. Gov Tech, www.tech.gov.sg/files/digital-transformation/dgb_summary_june2018.pdf

256. AI Singapore, www.aisingapore.org
257. SCMP, www.scmp.com/week-asia/business/article/2116290/why-tech-giants-see-singapore-next-artificial-intelligence-hub
258. Margaret Law, www.margaretlaw.com.sg/Article-An%20Overview%20of%20the%20Personal%20Data%20Protection%20Act.pdf

259. Cyber Security Agency (CSA), www.csa.gov.sg/~/media/csa/documents/publications/singaporecybersecuritystrategy.pdf
260. Smart Nation Singapore, www.smartnation.sg/initiatives

261. AI Singapore, www.aisingapore.org
262. National Research Foundation (NRF), www.nrf.gov.sg/docs/default-source/modules/pressrelease/201705031442082191-press-release-(ai.pdf

263. Tracxn, https://tracxn.com/explore/Artificial-Intelligence-Startups-in-Singapore/
264. Tracxn, https://tracxn.com/explore/Artificial-Intelligence-Startups-in-China/
265. Tracxn, https://tracxn.com/explore/Artificial-Intelligence-Startups-in-Australia/

267. IMDA, www.imda.gov.sg/about/newsroom/media-releases/2018/inaugural-meeting-of-the-advisory-council-on-the-ethical-use-of-artificial-intelligence-and-data
268. PDPC, www.pdpc.gov.sg/Resources/Discussion-Paper-on-AI-and-Personal-Data

269. PDPC, www.pdpc.gov.sg/Resources/Model-AI-Gov
270. MAS, www.mas.gov.sg/News-and-Publications/Media-Releases/2018/MAS-introduces-new-FEAT-Principles-to-promote-responsible-use-of-AI-and-data-analytics.aspx
271. GOV Tech, www.tech.gov.sg/digital-government-blueprint/
272. CIO, www.cio.com/article/3292616/how-singapore-is-using-artificial-intelligence.html
274. AI for Development, https://ai4d.ai/wp-content/uploads/2019/05/ai-gov-readiness-report_v08.pdf
275. Singapore Government, www.gov.sg/~/sgpcmedia/media_releases/imda/press_release/P-20180605-1/attachment/Artificial%20Intelligence%20Governance%20and%20Ethics%20Initiatives.pdf
276. IMDA, www2.imda.gov.sg/programme-listing/digital-services-lab

277. IMDA, www.imda.gov.sg/itsc/technical-committees/artificial-intelligence-technical-committee-aitc
278. Medium, https://medium.com/politics-ai/an-overview-of-national-ai-strategies-2a70ec6edfd
279. Ibid.
280. Marketing Interactive, www.marketing-interactive.com/sg-advisory-council-formed-to-create-ethical-ai-and-data-guidelines

281. Channel News Asia, www.channelnewsasia.com/news/technology/singapore-driverless-vehicle-autonomous-national-standards-11190498
282. Smart Nation, www.smartnation.sg/what-is-smart-nation/initiatives/Transport/autonomous-vehicles

283. Open Gov Asia, www.opengovasia.com/singapore-ranks-second-globally-in-autonomous-vehicle-readiness-report/
284. CIO, www.cio.com/article/3292616/how-singapore-is-using-artificial-intelligence.html
285. IMDA, www.gov.sg/~/sgpcmedia/media_releases/imda/press_release/P-20190123-1/attachment/SINGAPORE%20RELEASES%20ASIAS%20FIRST%20MODEL%20AI%20GOVERNANCE%20

FRAMEWORK.pdf
286. Gov Tech, www.tech.gov.sg/About-Us/Smart-Nation-Fellowship-Programme

287. Cyber Security Agency, www.csa.gov.sg/news/publications/singapore-cybersecurity-strategy

290. Korea Communications Commission (KCC), https://kcc.go.kr/user.do?boardId=1101&page=A02030900&dc=K02030900

291. MSIT, https://english.msit.go.kr/cms/english/pl/policies2/__icsFiles/afieldfile/2017/07/20/Master%20Plan%20for%20the%20intelligent%20information%20society.pdf
292. Korean Government, www.4th-ir2018.co.kr/bbs/download.php?bo_table=reference_en&wr_id=6&no=1

294. NIA, www.nia.or.kr/common/board/Download.do?bcIdx=20861&cbIdx=44086&fileNo=2

296. CIFAR, www.cifar.ca/docs/default-source/ai-society/buildinganaiworld_eng.pdf
297. AFR, www.afr.com/technology/fears-australia-will-be-left-behind-with-minimal-ai-spending-20190116-h1a4oo
299. Samsung https://research.samsung.com/aicenter_seoul
300. Reuters, www.reuters.com/article/samsung-elec-phones/samsung-to-invest-22-bln-in-ai-by-2020-idUSL4N1XI5LD; Channel News Asia, www.channelnewsasia.com/news/technology/samsung-

plans-us-116-billion-investment-in-non-memory-chips-to-challenge-tsmc--qualcomm-11473786
301. ORF Online, www.orfonline.org/wp-content/uploads/2018/11/Ai_Book.pdf

303. ORF Online, www.orfonline.org/wp-content/uploads/2018/11/Ai_Book.pdf
304. Samsung https://research.samsung.com/aicenter_seoul
305. Reuters, www.reuters.com/article/samsung-elec-phones/samsung-to-invest-22-bln-in-ai-by-2020-idUSL4N1XI5LD; Channel News Asia, www.channelnewsasia.com/news/technology/samsung-

plans-us-116-billion-investment-in-non-memory-chips-to-challenge-tsmc--qualcomm-11473786
306. KAIST, www.kaist.ac.kr/_prog/_board/?code=ed_camnews&mode=V&no=86781&upr_ntt_no=86781&site_dvs_cd=en&menu_dvs_cd=060102

307. Forbes, www.forbes.com/sites/cognitiveworld/2018/09/07/is-south-korea-poised-to-be-a-leader-in-ai/#376cea35fa2f
308. KAS, www.kas.de/documents/252038/4521287/Comparison+of+National+Strategies+to+Promote+Artificial+Intelligence+Part+1.pdf/397fb700-0c6f-88b6-46be-2d50d7942b83?version=1.0&t=1

559816454483#page=46
311. Bangkok Post, www.bangkokpost.com/business/1778849/first-draft-for-ai-ethics-framed

312. NESDB, www.nesdb.go.th/nesdb_en/ewt_dl_link.php?nid=4345
313. Thailand 4.0, https://thaiembdc.org/thailand-4-0-2
314. Digital Park Thailand, http://digitalparkthailand.org
315. Nation Multimedia, www.nationmultimedia.com/news/business/EconomyAndTourism/30308330
317. NESDB, www.nesdb.go.th/nesdb_en/ewt_w3c/ewt_dl_link.php?nid=4345
318. Thailand 4.0, https://thaiembdc.org/thailand-4-0-2
319. BOI, www.boi.go.th/upload/04_Digital_Park_Thailand&EEC_13589.pdf
320. Office of the Prime Minister, www.nesdb.go.th/nesdb_en/ewt_w3c/ewt_dl_link.php?nid=4345
321. Office of the Prime Minister, www.nesdb.go.th/nesdb_en/ewt_w3c/ewt_dl_link.php?nid=4345
322. Nation Thailand, www.nationthailand.com/Economy/30331361
323. Bangkok Post, www.bangkokpost.com/business/1688972/artificial-intelligence-takes-centre-stage
325. Bangkok Post, www.bangkokpost.com/thailand/special-reports/1614790/powering-up-on-ai
326. Can ASEAN, www.canasean.com/100firms-in-thai-to-adopt-ai-in-2019/
327. Channel Asia, https://sg.channelasia.tech/article/663471/thailand-transforms-healthcare-through-microsoft-ai
328. Healthcare Weekly, https://healthcareweekly.com/khon-kaen-launches-innovative-smart-health-solutions
329. Gov Insider, https://govinsider.asia/digital-gov/exclusive-interview-minister-digital-economy-thailand
332. UN-ESCAP, www.unescap.org/news/google-and-united-nations-economic-and-social-commission-asia-and-pacific-host-asia-pacific-ai
333. Gov Insider, https://govinsider.asia/digital-gov/exclusive-thailand-education-inequality-teerakiat-jareonsettasin
334. Nation Multimedia, www.nationmultimedia.com/detail/national/30351928
335. Thailand 4.0, https://thaiembdc.org/2019/06/17/tech-giants-helping-to-train-thai-digital-workforce
336. Open Gov Asia, www.opengovasia.com/iot-institute-to-anchor-thailand-digital-valley

338. Bangkok Post, www.bangkokpost.com/tech/1719147/thailand-56th-in-ai-readiness-index
340. Xin Hua Net, www.xinhuanet.com/english/2018-12/06/c_137653469.htm
341. Monetary Authority of Singapore (MAS), www.mas.gov.sg/Singapore-Financial-Centre/Smart-Financial-Centre/FinTech-Regulatory-Sandbox.aspx
342. Open Data Barometer, https://opendatabarometer.org/4thedition/regional-snapshot/east-asia-pacific
343. Korea Times, www.koreatimes.co.kr/www/tech/2017/07/133_233592.html
344. The Dialogue, http://thedialogue.co/dialogue/wp-content/uploads/2018/12/Report-Intersection-of-AI-with-Cross-Border-Data-Flow-and-Privacy-2.pdf
345. Salesforce, www.salesforce.com/company/news-press/stories/2019/06/062419-w
346. The Financial Review, https://amp.afr.com/technology/fears-australia-will-be-left-behind-with-minimal-ai-spending-20190116-h1a4oo
347. The Diplomat, https://thediplomat.com/2016/09/closing-asias-digital-divide
348. World Bank, www.worldbank.org/en/publication/wdr2016
349. ITU, www.itu.int/en/itunews/Documents/2018/2018-01/2018_ITUNews01-en.pdf
350. The Verge, www.theverge.com/2017/12/5/16737224/global-ai-talent-shortfall-tencent-report
351. McKinsey, www.mckinsey.com/~/media/mckinsey/business%20functions/mckinsey%20analytics/our%20insights/the%20age%20of%20analytics%20competing%20in%20a%20data%20

driven%20world/mgi-the-age-of-analytics-full-report.ashx
352. Tech in Asia, www.techinasia.com/singapore-introduce-programming-lessons-schools-boost-economy
353. Nanyang Technological University, www.ntulily.org/wp-content/uploads/conference/Deep_Model_for_Dropout_Prediction_in_MOOCs_accepted.pdf
354. Salesforce, www.linkedin.com/pulse/future-work-asia-upskilling-ai-jess-o-reilly-whittaker-
355. McKinsey, www.mckinsey.com/mgi/overview/2017-in-review/automation-and-the-future-of-work/jobs-lost-jobs-gained-workforce-transitions-in-a-time-of-automation
356. TRPC, http://trpc.biz/wp-content/uploads/IIC.Singapore-AI.BriefingPaper.20Nov.pdf
357. Enterprise Innovation, www.enterpriseinnovation.net/article/ai-must-be-more-human-new-era-business-automation-1177145624
358. New York Times, www.nytimes.com/2018/05/10/technology/alexa-siri-hidden-command-audio-attacks.html
359. UN-ESCAP, www.unescap.org/events/iic-escap-forum-artificial-intelligence-economic-growth-and-social-good

© The International Institute of Communications, all rights reserved
Highlands House, 165 The Broadway, London SW19 1NE
Tel: +44 (0)20 8544 8076 Email: [email protected]
www.iicom.org

1IIC AI Report 2020

Uploaded by

Copyright:

Available Formats

1IIC AI Report 2020

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1IIC AI Report 2020

Uploaded by

Copyright:

Available Formats

Artificial Intelligence

I. Executive Summary: Creating and Seizing AI Opportunities 6

II. The Promise of AI: Solving Problems, Improving Lives 14

III. The Challenge of AI: Key Topics to Address 20

IV. The Policy Response to AI: Overview of AI Strategies 34

V. Conclusion: Design Principles for Effective AI Policies 75

International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region 3

The IIC has always addressed policy issues head

4 International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region

Learn more at www.iicom.org

For more information, please visit www.trpc.biz

International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region 5

6 International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region

A fast-evolving technology that can serve multiple purposes, artificial

International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region 7

Other examples include the engineering student

8 International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region

The growth of AI across sectors, societies, and economies brings

Strong infrastructural foundations that provide reliable, widespread, and affordable

International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region 9

Skills and human capital

For businesses, this is embodied by the challenge of adapting workers, processes,

10 International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region

Developing data-driven innovations such as AI in a safe, ethical, and sustainable manner

This variety of hard-to-define technologies and systems can sometimes make it

Ecosystem and entrepreneurship

International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region 11

12 International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region

Types of Artificial Intelligence

Source: Deloitte, www2.deloitte.com/fi/fi/pages/technology/articles/part1-artificial-intelligence-defined.html;

International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region 13

14 International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region

This is also true for AI, which, as the AI for Good

In APAC countries, many initiatives are already

International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region 15

The widespread digitisation of human activity generates the kind of

AI tracks mosquito populations to reduce the spread of dengue

Project Wolbachia – a joint project between the National Environment

AI-powered microscope automatically spots Malaria parasites

The EasyScan GO is an AI-enabled microscope developed to assist

AI programme automatically screens for diabetic eye disease

Google is working with the state-run Rajavithi Hospital to launch an

16 International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region

AI enhances security in urban centres (Malaysia)

Since 2017, MIMOS, Malaysia’s national research and development

AI module customises students’ learning experiences (Vietnam)

Vietnamese tech giant FPT developed VioEdu, an AI-powered online

AI improves customer service and enhances e-commerce (Indonesia)

Kata.ai, Indonesia’s most popular conversational AI platform, works with

International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region 17

Whether directly or indirectly, it is increasingly clear that no sector or

Distributed energy grids

AI can enhance the predictability of demand and supply for renewable

Smart agriculture and food systems

AI-augmented agriculture involves automated data collection and

Weather and climate science and forecasting

A field that was once limited by its requirement of high-performance,

18 International Institute of Communications – Artificial Intelligence in the Asia-Pacific Region

AI can analyse simulations and real-time data to predict weather events