Levels of AIX:

The Future of AI and

the Human Experience
Overview

Artificial Intelligence (AI) is impacting the world we live in. It can already be found
across the systems, services and devices that we use every day. And while there’s
been exciting advancements in AI’s applications, the AI community lacks a common
framework and language for discussing advancements across the variety of domains,
such as on the road, at home, at work, and in public spaces.

This is why LGE and Element AI have partnered to research and develop this framework,
proposing a shared definition for advancements in AI. Grounded in the imaginative work
of foresight and research into the cutting edge of applied AI science and engineering, the
framework consists of four clear levels. Each level represents a step-change in capability
that will allow AI-powered products and services to provide new benefits to users and
society. We believe the future of AI should be human-centric, and so have coined the
term AIX - Artificial Intelligence Experience, which we hope will sharpen the public
discourse around this important topic.

To make clear what these different levels might mean in Unlike previous waves of automation that reshaped human
practice, we have included four underlying dimensions spaces and workflows to integrate machines, we believe the
and also considered ethical implications for how AI must transformative potential of modern AI is to reverse this trend.
remain safe and trustworthy at every stage. By creating devices, systems, spaces, and infrastructure that
adapt to people and their needs, increasingly sophisticated
AI can help unlock a new wave of growth in productivity and
well-being.

This is a lofty vision, and we recognize that building human-centric

commercial AI solutions will require expertise and effort from the
entire AI ecosystem. We see this framework as a starting point
for that vision and look forward to input and partnership with
researchers and industry, as we continue to collectively shape
the future of human-centric AI.
Here are the four levels
and their defining traits:

Level 1 Level 2 Level 3 Level 4

Names Efficiency Personalization Reasoning Exploration

Definitions AI facilitates specific functions AI uses pattern learning AI uses causality learning to AI uses experimental learning
with systems and devices, making to recognize, optimize and understand the cause of certain to continuously improve, by
user interactions more efficient personalize functions in patterns and behaviours, this forming and testing hypotheses
and effective order to improve and simplify information is used to predict it uncovers new inferences,
interactions for users and promote positive outcomes seamlessly adding value to users'
for users lives and enabling a deeper affinity

Pervasiveness Familiar Common Universal Foundational

In Our Lives
Systems and devices that AI is optimizing most devices AI is everywhere and AI forms a core component of
utilize AI are appearing at the edge and most systems interconnected for the benefit the infrastructure for all devices
in user’s everyday lives through the cloud of all devices and systems and systems in society which share
and learn collectively

Environmental Perceives Recognizes Understands Explores

Awareness
Perceives specific, pre-defined Recognizes patterns and uses Understands the patterns and Seeks to test and validate
information and acts on it them to make better predictions to principles across systems in order the underlying conditions of
accordingly to increase its increase relevance for users to meet predefined missions. Uses a situation by analyzing data
efficiency reasoning to respond to new froma broader set of external
situations by applying unique sources to inform its inferences
approaches

Collaboration Independent Connects Coordinates Orchestrates

Works alone or relays commands Connects with other devices within Understands the larger Identifies gaps in data and user
from one system to another a user-controlled system so that interconnected system and the understanding then orchestrates
the user can use one device to function of different devices and across internal and external
control others shares learning outcomes to systems to find and apply new
achieve a broader mission knowledge as it scrutinizes and
optimizes its hypotheses

User Agent Assistant Companion Sage

Understanding
Perceives user inputs and Recognizes and distinguishes Interprets the user’s mood from Understands how to influence
logs past inputs users and their unique contextual understanding of users - enabling them to trust
behaviours and preferences multiple data points and reasons new information and approaches
about social relations to predict by providing evidence, nudging
and support how users will interact behaviours in service of a broader
purpose

Autonomy Task-oriented Goal-oriented Mission-focused Purpose-driven

One-off actions Multiple actions Long-term actions Exploratory actions

Can execute specific commands Works out various options for Understands users and its Using local context and external
within specific parameters to achieving a given goal and environment in order to predict, sources of knowledge, it balances
achieve a specific task presents them to the user for recommend and execute solutions users’ competing needs and
selection or is pre-programmed to to assigned missions interests and is able to take
efficiently meet the desired goal creative approaches to influence
user behaviours, whilst in service
of the user’s higher purpose
Level 1 Level 1 Level 1

Preset with a grocery list, the Controlled by the household AI

smart fridge automatically orders that is programmed to conserve
groceries when items get low. water, the sprinkler senses there
was rain recently and so foregoes
watering the lawn.

Level 1

Level 2

With the help of a laundry robot,

clothes are picked from the floor,
sorted and washed according to
desired settings.

Level 2
Level 2 Level 2

Recognizing that the user has a business A smart mirror recognizes the user
trip in their calendar, the AI displays and displays biometric data taken
relevant trip info, suggesting a route to from wearable and environmental
the airport based on real-time road health sensors while reminding
construction updates. them to take their medicine.

Level 2

Level 3

Sensors on street lights monitor road

conditions, coordinating with individual
cars and triggering city-wide
traffic-controlling measures.

© 2019, LGE and Element AI www.lge.com www.elementai.com

Level 3 Level 3 Level 3

Based on data from wearable The weather forecast calls for snow.
health sensors, the AI makes diet The AI alerts the family to dress
and exercise recommendations to warm, preheats the oven and
help the user lose weight ahead of orders ingredients to prepare
an upcoming beach holiday. their favourite hot meal.

Level 3

The car’s AI is an extension of the home

and knows that the user is running late,
suggests altering the usual route to
ensure that an appointment is not missed
and provides a calming environment.

Level 4
Level 3

New research shows that reducing

screen-time in the evening will result
in improved sleep. The household AI
suggests a series of changes to the
family’s bedtime routine and
The AI senses the user is stressed about monitors for positive results.
a job interview in their calendar. It offers
to help them prepare by creating interview
questions and providing feedback, while
also setting the car route, the alarm and
suggesting wardrobe options.

A car’s AI interfaces with the smart city

to experiment with different routes,
departure times and driving speeds,
optimizing journeys based on daily user
objectives and other goals such as fuel
efficiency or journey time.

© 2019, LGE and Element AI www.lge.com www.elementai.com