798-Article Text-2442-1-10-20230109

Journal of Public Affairs and Development
Vol. 8: 33-65 (2021), ISSN 2718-9228
Adoption of Artificial Intelligence (AI)

in Local Governments: An Exploratory
Study on the Attitudes and Perceptions
of Officials in a Municipal Government
in the Philippines
CHARMAINE B. DISTOR1*, ODKHUU KHALTAR2,
and M. JAE MOON2,3
ABSTRACT. Emerging technologies like artificial intelligence (AI) have

been instrumental in transforming governments in recent years, which
is why several agencies worldwide have integrated them into their
governance strategies. One of the countries that have paid attention to the
potential of AI is the Philippines, which launched its national AI roadmap
in 2021. This study investigated the perceived acceptance and adoption
of AI in the Municipality of Carmona located in the Province of Cavite.
Following the combined constructs from the Technology Acceptance
Model (TAM) and Unified Theory of Acceptance and Use of Technology
(UTAUT), perception data were gathered from among Carmona’s local
executive and department administrators. The study found moderate to
high levels of positive attitude and behavioral intention toward the use of
AI in government operations. Moreover, correlation analysis revealed that
all acceptance and adoption factors had positive moderate to very strong
relationships with the attitude and behavioral intent to use AI, except
for the anxiety factor, which only had significantly negative moderate
to very weak correlation with attitude toward AI. As component of the
sequential mixed-method design, the thematic analysis provided context
to the findings. This paper contributes to the growing literature on the
governance, acceptance, and adoption of AI in governments.
Keywords: artificial intelligence, local government, TAM, UTAUT,

Philippines
________________________
1
United Nations University Operating Unit on Policy-Driven Electronic Governance
(UNU-EGOV), Guimarães, Portugal, https://orcid.org/0000-0001-6692-8854
2
Department of Public Administration, Yonsei University, Seoul, South Korea
3
https://orcid.org/0000-0002-6499-5700
*Corresponding author: [email protected]
Copyright 2023, the Authors. Published by the UPLB College of Public Affairs and Development.
This is an open access article licensed under Attribution-NonCommercial-ShareAlike 4.0 International
(http://creativecommons.org/licenses/by-nc-sa/4.0/).
34 Journal of Public Affairs and Development
Vol. 8: 33-65 (2021), ISSN 2718-9228
INTRODUCTION
Technology is ever evolving, and these changes cause some

disruptions to the way humans live. Technology experts have coined this
phenomenon as disruptive technology. Girasa (2020a) described the stages
of disruptive technology in two models. The first model consists of three
stages: paralysis, reaction, and transformation, whereas the second model
is composed of five stages: confusion, repudiation, shaming, acceptance,
and forgetting. Both models depict how disruptive technologies cause
organizations, including the public sector, to freeze at the initial stage
to observe and study the new technology. After a series of positive and
negative reactions, the organization, along with its stakeholders (e.g.,
citizens), learns how to accept and adopt the technology. Through the
years, various disruptive technologies have emerged (Girasa, 2020a) such
as cloud computing, blockchain, the Internet of Things, and the focus of
this paper – artificial intelligence (AI).
Various governments have started integrating AI into their

strategies (United Nations [UN], 2020; Girasa, 2020b; International
Telecommunication Union [ITU], 2017a). However, others still have some
reservations, especially in developing countries such as the Philippines.
The Philippine government can consider Mehr’s recommended strategies
(2017; Figure 1) in future developments of the recently developed national
AI roadmap. In addition, this study offers insights not only on the status
of the local governments’ ICT infrastructure but also on their personnel’s
attitude toward disruptive technologies such as AI, which could affect the
success of AI’s implementation across the country.
This paper addresses the research question: how do Philippine

local government officials perceive the use of AI for e-government
initiatives? Using the Technology Acceptance Model (TAM) and Unified
Theory of Acceptance and Use of Technology (UTAUT) frameworks
(Davis, 1989; Venkatesh et al., 2003), this research explores the case of
the Municipality of Carmona, Cavite in the Philippines through a mixed
methods approach to identify the perceived acceptance and adoption of
AI in the public sector.
AI Applications and Risks in Digital Government
AI is one of the most common disruptive technologies, which is

known for its ability to imitate human innate intelligence and behaviors
(Al-Mushayt, 2019). Computer scientist John McCarthy first introduced
Adoption of Artificial Intelligence (AI) in Local Governments: 35
An Exploratory Study on the Attitudes and Perceptions
of Carmona’s Government Officials in the Philippines
Figure 1
Recommended strategies for navigating AI in government (Mehr, 2017)
Make AI a part of a
goals-based, Build upon
citizen-centric Get citizen input
existing resources
program
Be data-prepared Mitigate ethical Augment

and tread carefully risks and avoid employees, do not
with privacy AI decision making replace them
the concept of AI in the 1950s. However, the logic behind AI has already
been utilized centuries ago. McCarthy’s AI became more prominent
in the present setting with the emergence of innovations such as Alan
Turing’s machine, which used mathematics to test its capability to play a
game of logic and strategy – chess (Girasa, 2020a).
Fast-forward to the 21st century, AI is becoming more relevant

in transforming institutions and societies. According to the 2019 report
by the Stanford Human-Centered Artificial Intelligence (Stanford
HCAI), the generation of AI-related jobs in the USA alone increased
by 0.5% from 2012 to 2019. AI-related startups also received increasing
investments globally with an annual increase of 48% since 2010. Gross
domestic products globally are also estimated to increase by around 26%
in 2030 due to AI alone (Kaya, 2019). Aside from the evident benefits of AI,
especially in the private sector, national governments have already started
integrating AI in their e-government initiatives as part of transitioning
toward digital governance. Some of the early adopters include South
Korea with its Intelligent Government Master Plan, the United Arab
Emirates’ Strategy for AI, and the European Union’s Digital Europe
Programme (UN, 2020). AI is also evident in smart traffic management
in countries like the USA and China (ITU, 2017a). The USA uses AI in
measuring performance in government agencies (Girasa, 2020b). AI also
emerged as an important technology in improving health services such
as Croatia’s virtual doctor, which acts as a medical assistant that answers
inquiries (UN, 2020).
Vol. 8: 33-65 (2021), ISSN 2718-9228
Figure 2
Artificial intelligence’s branches (Girasa, 2020a)
The field of AI is broad (Figure 2), and each aspect can help
improve public service delivery and the internal processes in the
government. For instance, natural language processing can be utilized
for creating chatbots (Oxford Insights, 2020). Machine learning can also
be used in generating legal proceedings predictions (Girasa, 2020a).
Dickson (2017) categorized AI into three types. First, artificial

narrow intelligence which can only perform one task at a time such as
the legal proceedings predictions. Second, artificial general intelligence
which is perceived as a strong AI for being able to mimic some of the
abilities of the human brain albeit not doing more complex tasks such as
reasoning. And third, the artificial super intelligence which is deemed as
a futuristic version of AI that can do everything a human brain can, and
even surpass it.
Similar to other e-government efforts, AI has its fair share of

advantages, risks, and challenges. Many of the news articles as well as
researches on AI in the past years covered AI ethics and data privacy
(Stanford HCAI, 2019). Some of the emerging malicious uses of AI
include massive surveillance and even the DeepFake technology that
contributes to fake news proliferation (Oxford Insights, 2020). A study on
AI’s use in medical treatments showed that only half of the respondents
believe in the technology’s ability and resist the idea of using AI for their
treatments (Carrasco et al., 2019).
Mehr’s (2017) study laid out the six common challenges faced
by governments with AI that are very evident in developing countries
such as the Philippines. These include: (1) resource allocation, (2) large
datasets, (3) AI expert shortage, (4) predictable scenario, (5) repetitive
procedures, and (6) data diversity.
Current State of AI Applications and Digital Government
Based on the 2022 E-government Survey of the United Nations,

the Philippines ranked 89th on the E-government Development Index
(EGDI), which is twelve steps lower than its 2020 ranking (UN, 2020
& 2022). Although the Philippine information and communications
technology (ICT) infrastructure is still developing unlike other ASEAN
countries, which are already advanced, the latter reflects surprising
progress. For instance, the penetration of mobile services in the country
has exceeded the 100% mark (ITU, 2017b).
However, the Philippine government still needs to focus on

improving its Internet and broadband services, especially in terms of
accessibility and bandwidth. Only 60% of the population have Internet
access and around 3% have fixed broadband subscriptions (UN, 2020).
In terms of bandwidth per Internet user, the Philippines exhibits slow
performance compared to the Asia-Pacific (0.384 Mbps) and global (0.596
Mbps) averages with only 0.3472 Mbps (ITU, 2017b).
Nevertheless, some improvements in the Philippine ICT

infrastructure have been observed during the pandemic. A report
by the Philippine Daily Inquirer (Baclig, 2021) showed that major
telecommunication companies in the country recorded an increasing
number of subscribers as people transitioned to study and work from
home set-ups. The same report also cited Ookla’s Speedtest Global Index
that there has been an enhancement in the country’s Internet speed as it
moved 15 places higher from its last ranking.
In terms of AI adoption, the Philippines is tagged as one of the

lagging countries based on the AI Index 2019. It pales in comparison
to its neighboring countries such as Vietnam, Thailand, Indonesia, and
Malaysia, which are named or tagged as emerging and to Singapore,
which is in the cutting edge category (Stanford HCAI, 2019). But when
it comes to the Government AI Readiness Index (GAIRI) 2020 by the
Oxford Insights (Figure 3), the Philippines is ranked 74th, which is higher
than Vietnam (75th), Laos (127th), Cambodia (128th), and Myanmar (131st).
The report explains that one of the reasons behind the low ranking of
ASEAN countries such as the Philippines in GAIRI 2020 is the existing
digital divide and the still developing ICT and AI infrastructure in the
region.
Unlike countries with established AI strategies and policies,

the Philippine government is still in its formation stage. In April 2019,
Department of Trade and Industry (DTI) Secretary Ramon Lopez said in
Vol. 8: 33-65 (2021), ISSN 2718-9228
Figure 3
The Philippines and select Asian countries’ Government AI Readiness
Index (Oxford Insights, 2020)
an interview that the country seeks to be known as a“global AI hub”. He

further explained that the country has more than 100,000 graduates from
the field of science, technology, engineering, and mathematics (STEM)
annually who can be mentored in the field of AI by mentors like Dado
Banatao, a Filipino technopreneur with a thriving career in the Silicon
Valley (Dumlao-Abadilla, 2019). The Philippine government launched
the AI roadmap on 5 May 2021, which laid down several government
strategies for maximizing AI’s potentials such as through the creation of
the National Center for AI Research or N-CAIR (DTI, 2021).
As early as 2017, Filipino policymakers also showed interest

in AI. Senator Paolo Benigno “Bam” Aquino introduced the Senate
Resolution No. 344 or the “Plans and Initiatives of the Government on
the Developments in Artificial Intelligence”(2017). In 2019, Senator Ralph
Recto also introduced one controversial AI issue through the Senate
Resolution No. 188 or the “Proliferation of AI-Synthesized Audiovisual
Materials, known as DeepFakes” (2019). Despite the promises of both
introductory resolutions, both are still pending in the science and
technology committee of the Congress.
Model and Theory on Technology Acceptance
This study applied the TAM and UTAUT because they are two
of the most used technology acceptance model and theory, especially in
relation to using AI, and they are highly compatible with one another for
sharing the same determinants-attitude-intention thread.
AI alongside other emerging and disruptive technologies are

gaining the attention of governments globally mainly for their economic
and operational benefits (Stanford HCAI, 2019; Kaya, 2019; UN, 2020).
However, little is still known for the other drivers of acceptance and
adoption of AI. While a wide range of studies have applied TAM and
UTAUT to AI, examples of their application in the government are still
scant. Hwang and Nam (2020) studied how the socio-demographics of
Korean citizens affect their AI acceptance and found that higher education
and high trust in government positively affected their acceptance.
Another study focused more on the institutional use of AI such as
Arizona’s COPLINK, a knowledge management system for law enforcers
(Lin et al., 2004). Through the said study, which used TAM alongside
the Theory of Planned Behavior, law enforcers’ perceived usefulness
was shown to be the most crucial factor affecting their acceptance to
use COPLINK. Another case study from the United Arab Emirates
government investigated AI in the health sector and found that several
factors were integral to increasing medical practitioners’ perception of
AI’s functionality and practicality—the organization, management, and
the existing ICT infrastructure (Alhashmi et al., 2019).
Besides these studies, there is still an existing research gap on

what motivates governments to accept and adopt AI in their governance
strategies especially from the perspective of the main decisionmakers.
Is it driven by the hype of the technology? Or is it driven by the actual
need of the government agencies and their constituents? To provide a
deeper understanding of the drivers of AI acceptance and adoption in
governments, this study focused on the case of Philippine government
officials in one municipal government drawing from the constructs of
the Technology Acceptance Model (TAM) and the Unified Theory of
Acceptance and Use of Technology (UTAUT). The researchers presumed
a sequential relationship between a combination of adoption factors
and two dependent variables: attitude toward using AI and behavioral
intention to use AI.
Technology Acceptance Model (TAM). Davis (1989) argued

that one’s acceptance of new technology is determined by his or
her behavioral intention. Two main determinants are responsible for
behavioral intention. First, he defined perceived usefulness as the user’s
extent to see the technology as beneficial in enhancing one’s performance.
AI-related studies in different fields such as medicine have supported
the notion that the perceived usefulness has a significant relationship
with the attitude and behavioral intention to implement the behavior
(Alhashmi et al., 2019). On the other hand, Davis used the perceived ease
Vol. 8: 33-65 (2021), ISSN 2718-9228
of use to refer to the extent of user’s effort in using the technology. One
paper studying China’s telecommunications sector found that Chinese
users preferred simplicity and user-friendliness compared to high-tech
yet complicated services (Rahi et al., 2018). Alhashmi et al.’s review
of related literature (2019) argued that perceived ease of use can have
indirect and direct impacts on both the attitude and behavioral intention.
Perceived usefulness is a mediating factor of perceived ease of use toward
behavioral intention, yet perceived ease of use has a significant influence
over behavioral intention beyond the effect of perceived usefulness (Davis,
1989). Moreover, Davis together with Bagozzi and Warshaw explained in
their 1992 paper that in TAM, it is postulated that one’s attitude toward
a technology can also affect behavioral intention. This stemmed out of
TAM’s theoretical inspiration, the theory of reasoned action.
Unified Theory of Acceptance and Use of Technology

(UTAUT). Proponents of the UTAUT, Venkatesh and his colleagues
(2003), have aligned arguments with Davis (1989) but use the terms
performance expectancy and effort expectancy instead as key predictors
of behavioral intention to use new technologies. In particular, they
agreed with the definition (Davis, 1989; Davis et al., 1989) that associated
usefulness with relative advantage and extrinsic motivation.
Another key factor that they added is social influence, defined

as the extent of the user’s environment and social circle affecting his/
her technology usage. In Venkatesh’s recent article (2021), he argued
the timeliness of UTAUT in the age of AI. The UTAUT proponents
also refer to facilitating conditions as the infrastructure supporting the
technology. However, beyond the infrastructure, which usually translates
to software and hardware, studies have shown how facilitating conditions
related to the organization’s culture and practices such as information
dissemination may affect the adoption of the technology (Gruzd et al.,
2012). In UTAUT, self-efficacy is defined as the user’s belief in his or her
capacity to use the technology. Venkatesh (2021) further explained that
individual characteristics such as self-efficacy are very influential toward
technology adoption and those who are more “risk-seeking, tolerant of
uncertainty, and with a desire to learn” have higher chances of accepting
and adopting the technology. Meanwhile, Venkatesh et al. (2003) defined
anxiety as the user’s apprehensions toward the technology. Although
some studies view this determinant as non-significant toward technology
acceptance and adoption, Gruzd et al.’s (2012) study showed how some
forms of anxieties such as time constraints and information overload
hinder academicians from using social media for their research. Lastly,
the behavioral intention in UTAUT refers to the user’s intent to use the
technology.
Conceptual Framework
Guided by the TAM and UTAUT, a case study was conducted to

identify the perceived acceptance and adoption of AI in the Philippine
government. AI acceptance is determined by two factors such as
attitude toward using artificial intelligence and behavioral intention to
use artificial intelligence. On the other hand, six adoption factors were
considered, namely: perceived usefulness, perceived ease of use, social
influence, facilitation conditions, self-efficacy, and anxiety (Figure 4).
Figure 4
Conceptual framework combining TAM and UTAUT
METHODOLOGY
Research Location
The case study was conducted in the Municipality of Carmona

in the Province of Cavite in the Philippines. The local government unit
(LGU) of Carmona has a pending line-up of AI projects such as the AI-
enabled chatbot for the local government’s official Facebook page and
the use of facial recognition technology for barangay or village-level
registry system. The national government has also recognized Carmona
LGU as a Center for Excellence for its ICT programs. Moreover, Carmona
has received the most coveted Seal of Good Local Governance for five
years in a row where ICT innovations play a big role.
Vol. 8: 33-65 (2021), ISSN 2718-9228
Mixed Methods Approach
This study used a mixed methods approach to capture the

perceptions of AI adoption in local governments. While most studies
utilizing the TAM and UTAUT only used the quantitative approach
(Lee & Baskerville, 2003), recent researches also deemed it necessary to
integrate a qualitative lens in understanding technology acceptance and
adoption (Jayaseelan et al., 2020; Hanif & Lallie, 2021; Williams et al.,
2021; Zhu & Zhang, 2021). Specifically, this study used the explanatory
sequential mixed methods design, which constitutes two phases of data
gathering: the quantitative data is collected and analyzed first followed
by the collection of qualitative data. As defined by Creswell and Creswell
(2018), the intent of this type of mixed methods design “is to have the
qualitative data help explain in more detail the initial quantitative results”
(p. 304).”
The authors deployed a self-administered survey to all 20

department heads of the Carmona LGU after orienting them about AI
and its implications in their respective offices. The survey questionnaire
included three to four statements for each of the six adoption factors and
two dependent variables adapted from earlier TAM and UTAUT studies.
Responses were weighted through a semantic Likert scale (1 or “strongly
disagree” to 7 or “strongly agree”). The responses were descriptively
analyzed and subjected to Pearson correlation using STATA 13.
Thematic analysis was conducted to explain and contextualize

the quantitative findings (Braun & Clarke, 2006). Selected department
heads and the municipal mayor were interviewed in-depth to generate
needed information. To ensure that all department heads were well-
represented in the interviews, the sample of respondents were selected
based on the Beresford Research’s age range by generation (2021). Only
the Millennials (1981 – 1996), Generation X (1965 – 1980), and Boomer
II (1955 – 1964) were actually represented in the pool of the Carmona
LGU’s department heads. Aside from the age, this study also ensured
that all department clusters were represented – administrative/legislative,
social, infrastructure, and finance (For the demographic information of
the survey and in-depth interview respondents, refer to Appendix 1).
In total, 21 open-ended questions around the concepts

presented in the TAM and UTAUT questionnaires were fielded to the
interviewees with follow-up questions for probing. The duration of the
interviews varied between 21 and 73 minutes. All interview sessions were
conducted and recorded using the Zoom video conferencing tool. The
interviews were manually transcribed, translated, and coded to generate
salient codes and subcodes.
RESULTS AND DISCUSSION
Descriptive Statistics and Correlation Analysis
Table 1 shows the overall mean scores of all AI acceptance and

adoption factors among the surveyed department heads in the Carmona
LGU (for the descriptive statistics of all questionnaire statements, refer to
Appendix 2). The department heads mostly agreed with the statements
describing four variables with the highest means, namely: perceived
usefulness (5.75), perceived ease of use (5.43), social influence (5.63), and
attitude toward using AI (5.96).
Table 1
Overall mean scores of AI acceptance and adoption factors
Variable Overall mean score

Behavioral intention to use AI 5.08
Attitude toward using AI 5.96
Perceived usefulness 5.75
Perceived ease of use 5.43
Social influence 5.63

Facilitating conditions 5.13
Self-efficacy 5.25
Anxiety 3.63
Meanwhile, facilitating conditions received an overall mean of

5.13, self-efficacy was 5.25, and behavioral intention was 5.08. These
mean scores demonstrate that the respondents were ambivalent to the
statements related to facilitating conditions, self-efficacy, and behavioral
intention. Only anxiety received a low yet neutral overall mean score
(3.63).

In terms of the correlation of the AI acceptance and adoption
factors with behavioral intention (Table 2), perceived ease of use (r
= 0.8501) and perceived usefulness (r = 0.8373) showed the strongest
correlations. This means that the perception of the department heads
of Carmona LGU on how easy it is to use an AI intervention and how
they see AI’s usefulness in their departments’ functions were strongly
associated with higher behavioral intention to use the AI tool. Attitude
toward using AI also exhibited a strong correlation with behavioral
Vol. 8: 33-65 (2021), ISSN 2718-9228
Table 2
Correlation of AI acceptance and adoption factors
Variable BI AT PU PE SI FC SE AX
BI 1.0000
AT 0.7224* 1.0000
PU 0.8373* 0.7476* 1.0000
PE 0.8501* 0.7722* 0.8757* 1.0000
SI 0.6930* 0.8136* 0.8170* 0.7586* 1.0000
FC 0.6895* 0.5007** 0.5905* 0.6453* 0.5484** 1.0000
SE 0.5647* 0.4073*** 0.5979* 0.6990* 0.5202** 0.5112** 1.0000
AX -0.3002 -0.5490** -0.5445** -0.5360** -0.6275* -0.1653 -0.3617 1.0000
Notes: *p<0.01, **p<0.05, ***p<0.10 (2-tailed)
PE: Perceived ease of use AT: Attitude toward using AI

PU: Perceived usefulness AX: Anxiety
SI: Social influence BI: Behavioral intention to use AI
SE: Self-efficacy FC: Facilitating conditions
intention (r = 0.7224), which can be construed as a positive attitude of

the department heads toward AI leading to higher behavioral intention
to use the AI interventions. Whereas, social influence (r = 0.6930) and
facilitating conditions (r = 0.6895) also obtained strong correlations while
self-efficacy (r = 0.5647) received a moderate correlation with behavioral
condition.
Meanwhile, for the correlation of attitude toward using AI, social

influence showed the strongest relationship (r = 0.8136). This indicates
that the people who have influence over the department heads such
as the mayor, their colleagues, and even their loved ones may have an
effect on the latters’ attitude toward using AI in their respective offices.
Perceived usefulness (r = 0.7476) and perceived ease of use (r = 0.7722)
also demonstrated strong correlations with attitude toward using AI.
Facilitating conditions (r = 0.5007) and self-efficacy (r = 0.4073), on the
other hand, showed moderate correlations with attitude toward using
AI. Anxiety (r = -0.5490) is the lone variable which exhibited significant
negative correlation with attitude toward using AI.
Several AI acceptance and adoption factors have also shown very

strong relationships with one another such as perceived usefulness and
perceived ease of use, which obtained the highest correlation (r = 0.8757).
This indicates that when the department heads see that the AI tool has a
user-friendly interface and it is easy to navigate, they may most likely have
a higher perception that the AI tool can make their department’s duties
more efficient. Social influence likewise showed a very strong correlation
with perceived usefulness (r = 0.8170), which implies that the usefulness
of AI is affected by the social environment of department heads (e.g.,
their colleagues, the mayor, and their families). A practical example of this
is if ever the municipal mayor does not favor the AI intervention in their
department, the department head would most likely see AI as useless.
On the other hand, most variables also exhibited a strong relationship
with perceived ease of use. For instance, a department head’s perceived
ease of use of AI in their office’s programs would go higher with their
individual self-efficacy or self-reliance in using the AI intervention (r =
0.6990).

It can also be observed that some anxiety relationships,
particularly with facilitating conditions (r = -0.1653), self-efficacy (r =
-0.3617), and most importantly behavioral intention (r = -0.3002), were
not significant on any of the given significant levels (1%, 5%, 10%). This
supports the claims made by Venkatesh and colleagues (2003) when they
formulated UTAUT wherein anxiety and self-efficacy were deemed to
Vol. 8: 33-65 (2021), ISSN 2718-9228
be indirect drivers of behavioral intention. Furthermore, all the observed

negative relationships were only exclusive to anxiety. This finding is a
good sign since this implies that the worries of the department heads
about AI were limited and may not have any significant effect with the
other variables.
To further explain the correlation findings, the in-depth

interviews were subjected to a thematic analysis. Findings are discussed
in the next section.
Qualitative Findings
Previous Use of AI. The respondents had prior experience in

using AI-enabled technologies. Most of them used AI in the form of
digital voice assistants such as Amazon’s Alexa and Apple’s Siri using
their personal smartphones. Another AI application that the respondents
were accustomed to using for personal use is the chatbot available on
e-commerce apps and websites. In terms of work-related use of AI, the IT
department head was very knowledgeable because of his technical and
educational background. “When I was taking my master’s, we developed
a loan system using artificial neural networks,” the IT department head
shared. The health department head also shared that he was able to use
AI programs that their national government counterpart cascaded down
to local governments, especially during the COVID-19 pandemic. Aside
from work-related AI experience in the public sector, the accounting
department head also said that he was able to use an AI system in the
past when he worked as an accountant for multinational firms.
Perceived Usefulness. Most of the themes that emerged from

the interviews are related to increasing the performance and efficiency
of the local government services. Some of these include reduced
errors, enhanced accuracy, and increased time efficiency. The planning
department head also explained that since their department oversees the
local census by using the community-based monitoring system (CBMS)
technology, they handle a huge amount of data, and sometimes, the
storage and retrieval of the data are tricky and time-consuming.“We face
some backlogs in the current system of using CBMS, and sometimes, when
some people request data from our department, it takes time to generate them.
I think with AI, we can better manage our data. Uploading, processing, and
linking data would much be easier too,” she added.
The municipal administrator and IT department head both

shared that AI can be useful both for internal and external users citing
that it can upgrade the skills of civil servants and enable them to focus
more on high-level and non-routinary tasks. Further, most respondents
agreed that AI can make transactions for local government clients more
convenient because they would not have to travel to the municipal hall
just to avail of some services. The municipal mayor said that AI can help
him and the local government in general in decision making, but he
stressed that “AI should only be a guide on decisions and not entirely be the
one to have the final say.”
Perceived Ease of Use. Majority of the respondents agreed that

the local government employees and their fellow department heads can
easily learn new technology like AI. The accounting department head
explained that employees will be motivated to learn AI since it is an
additional skill set for themselves. The municipal mayor also shared the
same sentiment and emphasized that it is also a matter of compliance
with the vision of Carmona LGU, and if the employees will not adjust
easily, they and their departments will be left behind in the digital
transformation. “I believe in the willingness of the employees of our local
government. I think, as long as there’s passion to improve, dedication to serve,
you will never get tired of learning including learning new technologies,” the
mayor said.
However, most of the respondents agreed that the ease of use

might be heavily affected by the age of the users. Both the professional
bureaucrats – the accounting department head (a certified public
accountant) and the health department head (a physician) shared
that there can be some bottlenecks in learning new technology. “Older
healthcare workers are not that techy. For younger healthcare workers, they
might be digital natives, but based on my observation, they also receive the
biggest workload. They might also find it hard to insert studying AI in the midst
of their jobs,” the health department head said.
The IT department head highlighted that most of the time, people

especially the local government’s constituents focused too much on how
“high-tech” an innovation is, but he clarified that, “we must actually focus
not on how high-tech an innovation like AI can be, but rather on how easy it is
to understand and use it. I also think that a simple technology can increase its
acceptance and adoption.”
Vol. 8: 33-65 (2021), ISSN 2718-9228
Social Influence. In terms of the influential personalities on the

future implementation of AI in Carmona LGU, all of them mentioned
the mayor first including the local chief executive himself. “I definitely can
say that mayors like me have the say on innovations like AI, so yes, it’s me,”
he shared. The accounting department head affirmed that the current
mayor is very supportive of innovations and urges all department heads
to suggest plans to him. The IT department head supported this opinion
and even added that the mayor “is also a very effective and maybe the best
champion of new technologies like AI. We know that he is already old, yet he
still supports innovation. And I think, if I also back him up when he vouches
for technologies like AI, his credibility on the subject increases, and citizens will
most likely be convinced to further use AI.”
The health department head said that since they have direct
contact with the citizens as healthcare practitioners, they can also heavily
influence their patients as “champions of innovations.” Meanwhile, the
municipal administrator shared that for her, other like-minded and
innovation-oriented local governments such as Valenzuela City and Pasig
City in the National Capital Region can influence the Carmona LGU
to adopt AI and other technologies or to “embrace technology beyond
politics.”
Facilitating Conditions. Both the answers of the mayor and IT

department head are in sync in terms of how ready the Carmona LGU is
in implementing AI.“I personally think that AI is very costly, but I think we
can start implementing it phase by phase,” the mayor said.
The IT department head added,“I also think that it depends on the

level of AI that will be deployed. It is easy if it is software-related, and I think
we can implement it right away. But if it’s hardware or something complex, I
will be honest that our department’s load can’t take it. And maybe, if there’s
a need to implement it right away, we can outsource for an external service
provider.”
According to the planning department head, budget is very

essential in implementing AI, but building the capacity of the employees
and the citizens are equally as important. The municipal administrator
has the same opinion and emphasized having pilot sites first before the
implementation so that resources will not be wasted.
Self-Efficacy. Table 1 shows that most department heads are

ambivalent with the statements used to measure self-efficacy. This can
be explained by the results of the in-depth interviews. All the respondents
preferred to be guided step-by-step when AI will be introduced to them

rather than rely on themselves when navigating AI for the first time. “I
think if I will only rely on myself and same for my staff, we will face a lot
of errors and maybe the implementation will be even slower because of those
errors,” the accounting department head imparted.
The health department head shared that their office cannot

afford to commit any mistakes because the health and even the lives of
their patients are at stake. According to her, “It’s really critical for social
services like us. And if ever we commit errors when we use the AI without any
expert’s help, it will definitely affect their trust not only on our department, but
also on new technologies in the future, and on the local government. So, I prefer
to be assisted step-by-step…slowly but surely.”
Anxiety. As shown in Table 2, anxiety had a significant negative

correlation with attitude toward using AI as well as other acceptance and
adoption factors like perceived usefulness, perceived ease of use, and
social influence.
Most of the themes of the issues raised by the respondents

were more on external factors. First, data privacy can be an issue. “We
experienced it when we launched our local contact tracing app last year for
COVID-19. Some concerned citizens even went to our office because of it…the
use of their data is not clearly stated on the app. It’s a lesson learned,” the IT
department head shared.
Second, the Carmona LGU is very responsive to its constituents,

and they need to devise an efficient method to obtain objective feedbacks
to improve future AI programs. “Although it’s a possible application of AI
from another department [social welfare and development department], I think
the citizens’ acceptance of AI might be affected if they were not selected by the
AI, for example in financial assistance programs, which have certain criteria. I
think we should be able to create an objective feedback mechanism for that,”the
health department head suggested.
The municipal administrator also mentioned another external

concern related to existing national laws, which for her are not “AI-
friendly”: “They should really revisit the policies because sometimes, even if
the local government wants to get things automated, for example we want to
send the financial assistance to our constituents by online means, the national
government will still look for documentation like a manually signed payroll of
the beneficiaries.”
Vol. 8: 33-65 (2021), ISSN 2718-9228
In terms of the internal anxieties, the accounting department

head mentioned that overdependence and excessive trust on AI can also
be an issue as the civil servants would be less motivated to check the
errors themselves.
Attitude Toward using AI. In Table 2, all variables have a

significant correlation with the attitude toward using AI, which had an
overall mean of 5.96 (Table 1). The interview findings likewise supported
the turnout of the exploratory survey. For instance, the accounting
department head said that he will most likely use AI since he perceives
AI to truly contribute to making their department’s functions easier. This
supports the strong correlation found between perceived usefulness and
attitude toward using AI. In addition, he shared that he is motivated
to accept AI wholeheartedly because the mayor encourages and urges
departments to innovate. This, in turn, supports the relationship between
attitude and social influence which acquired the highest correlation
coefficient (r=0.8136).
The other respondents, such as the planning department head,

shared the same views as the accounting department head who mentioned
that “since AI is perceived to make transactions faster, make the government
more transparent, and it increase citizen satisfaction, then definitely, we will
implement AI-enabled projects soon.” But even if there is high acceptance
of using AI among the department heads of the Carmona LGU, both the
health department head and the municipal administrator remarked that
the local government must still evaluate closely which services should
apply AI. They contended that there are still some public services that are
better delivered with human contact such as health services.
Behavioral Intention to Use AI. Similar to the attitude toward

using AI, the behavioral intention also received high agreements both
in the exploratory survey as well as in the interviews. The local chief
executive himself is very enthusiastic when asked about the timeline of
the local government to implement its AI-enabled programs.“I project it is
within the next two years. I already started discussing with the IT department
head and other heads. Some of the projects we have on the top of our head right
now include using facial recognition for law enforcement,” the mayor shared.
The planning department head also mentioned that since the

Carmona LGU’s vision in the coming years is to be a smart city, they
are already preparing the infrastructure such as a local data center.
She believes that this will be helpful when AI programs begin to be
implemented one by one. The health department head added that to
make Carmona AI-ready, other sectors of the community, especially the

firms and companies inside the municipality, should also start using AI
when the local government starts to launch AI programs.
Discussion
This study was exploratory and designed to draw inference

about causal relationships between the six adoption factors drawn from
TAM and UTAUT. Results of the correlation and thematic analysis reflect
consistency with current body of research on AI acceptance.
The correlation results align with observations in previous studies

on AI acceptance and adoption albeit in the private sector. For instance,
in Sohn and Kwon’s paper (2020) about the development of AI-based
intelligent products in South Korea like smart speakers, voice assistants,
and smart appliances, they found that perceived usefulness, perceived
ease of use, and attitude have positive moderate to strong correlations
with behavioral intention. Another case study from China focused on AI-
based medical diagnosis support systems (Fan et al., 2020) also showed
positive strong correlations of behavioral intention with effort expectancy
(or perceived ease of use in this paper), performance expectancy (or
perceived usefulness in this paper), as well as social influence. In Kim
et al.’s (2019) study of finance chatbots, they found a significant positive
relationship between behavioral intention and adoption factors such as
effort expectancy, performance expectancy, and social influence. Another
South Korean case study on healthcare AI (Kim, 2017) shared the same
trends with the addition of a significant negative relationship between
anxiety and behavioral intention.
As summarized in Table 3, perceived usefulness’ sub-themes

included performance improvement, convenience, and decision-making.
In Kessler and Martin’s (2017) study on the perceptions of potential users
of AI voice assistants, most of their focus group discussion participants
appreciated how voice assistants made people’s lives easier – or the
variable of perceived usefulness. Another qualitative study on AI
highlighted perceived usefulness through themes such as time efficiency
and diagnostic quality, among others (Buck et al., 2022). Meanwhile, this
study’s sub-themes for perceived ease of use are similar to previous AI
perception studies. For instance, technology simplicity appeared as a key
theme in the interviews, which also resonated with Kessler and Martin’s
finding (2017) that AI should “work without any hassle.” Age was also
one of the key sub-themes of perceived ease of use in this study, which
Vol. 8: 33-65 (2021), ISSN 2718-9228
Table 3
Summary of the thematic analysis of the interviews
Theme Sub-theme Supporting narrative

Perceived Performance improvement AI as a tool for increased performance, better time efficiency, less errors, and
usefulness less backlogs
Convenience AI as a tool that would bring convenience both for the civil servants and the
constituents
Decision-making AI should only be a guide on decisions
Perceived Age Younger civil servants would find it easier to learn AI compared to the older
ease of use ones
Additional skillset Learning AI is seen as an additional skillset and those who won’t adapt can be
left behind
Technology simplicity AI implementation should be focused on how easy it can be used rather than
on how high-tech it appears to be
Social Mayor as the main influencer Local chief executive is the critical proponent in influencing decisions on AI
influence implementation
Civil servants as champions of Civil servants especially the street-level bureaucrats can influence citizens on
innovations AI adoption
Like-minded local governments Other innovation-oriented local governments can be influencers
Facilitating Budget AI implementation is a costly project so the creation of pilot sites to avoid
conditions wasting resources
Capacity building Increase the AI-related capacity and skills of civil servants and constituents
Outsourcing Although the IT department can handle the creation of AI tools, outsourcing
service providers can be an option to hasten implementation
Table 3
Continued
Theme Sub-theme Supporting narrative

Self-efficacy Guided learning Preference to be guided step-by-step on using AI tools
Critical mistakes Social services can’t afford to make mistakes
Trust When civil servants are not knowledgeable on AI and commit mistakes,
this could affect the people’s trust in the local government and on future
technologies
Anxiety Data privacy Use of citizen data in AI initiatives should be considered
AI’s decisions Balance between AI’s objectiveness and citizen satisfaction
AI-friendly policies Some existing national policies inhibit the implementation of AI by the local
government
Excessive trust on AI Some civil servants may develop overdependence on AI and might be less
motivated to check errors themselves
Attitude Local government efficiency There is a positive attitude toward AI brought by perceived notion that AI
toward using would make department functions easier
AI Government transparency AI is perceived to make government processes more transparent
Citizen satisfaction AI is perceived to make transactions faster, which could also increase
citizen satisfaction
Behavioral Smart city Since Carmona is aligning itself to be a smart city, AI implementation can
intention to be a priority
use AI Holistic AI implementation AI should also be implemented across all sectors including businesses
Vol. 8: 33-65 (2021), ISSN 2718-9228
resonates with the arguments of the proponents of UTAUT that age is

a moderator between effort expectancy (or perceived ease of use) and
behavioral intention (Venkatesh et al., 2003).
For facilitating conditions, the department heads and the mayor

of Carmona LGU acknowledged that implementing AI projects is very
costly, which pointed to the sub-theme budget. This was also reflected
in the findings of Atwal et al. (2021) where finances emerged as an
important aspect of resources in the winery industry AI.
The Carmona LGU’s case affirmed the need for AI stakeholders

to upskill in AI such as those found in recent findings in AI for supply chain
management (Hasija & Esper, 2022). In terms of self-efficacy, one of the
important sub-themes that appeared in the interviews in the Carmona
LGU was trust. As pointed out by one of the department heads, a simple
mistake due to a lack of knowledge or skill in AI may compromise the
trust of citizens not only on AI but also on future technologies; it may
even jeopardize their trust on the local government.
Moreover, this study revealed several sub-themes pertaining to

anxiety. One of them is data privacy. This appeared in the narratives of
respondents from past studies such as the AI voice assistants who were
apprehensive that everything they say may be heard (Kessler & Martin,
2017). The Carmona LGU respondents were also anxious about the AI-
friendliness of existing policies, which would impede AI implementation.
The same concern was raised in one French AI case study (Atwal et
al., 2021) where the adoption of automated harvesting equipment was
not supported in the locality because local policies prescribed manual
harvesting for wine production.
As this study is one of the few literatures covering AI acceptance

and adoption in government, several unique sub-themes were also
observed. For social influence, this study showed more organizational
influencers for the citizens such as the mayor, civil servants, and other
like-minded local governments. Previous studies showed influences
driven by personal connections like family and friends. In terms of the
attitude toward using AI, positive attitudes were detected in the interviews
as the department heads and the mayor associated AI implementation
with local government efficiency and government transparency. Another
important attitude sub-theme is citizen satisfaction, which is supported
by past local government studies wherein local governments were usually
driven to do programs (e.g., AI implementation) because of increased
public satisfaction that affects election turnout (Hong, 2017). When it
comes to behavioral intention, the Carmona LGU is at an advantage
compared to other municipalities as there is political will backing up their

smart city positioning. Since they already have this renewed vision, the
integration of AI projects would not be too difficult especially since the
mayor is also a key social influence. Moreover, another theme mentioned
in the narratives for behavioral intention is the holistic implementation of
AI across the local government. The department heads emphasized that
AI adoption should not only be a concern of the local government but
also of other sectors such as companies in the municipality.
CONCLUSION AND RECOMMENDATIONS
This study offers insights on how Philippine local government

officials perceive the use of AI for e-government initiatives, using the case
of the Carmona LGU. Findings contribute to the growing literature on AI
acceptance and adoption, especially in the government setting, which is
still relatively sparse.
AI is a disruptive technology that provides several potentials that

can be utilized by governments to improve their services to the public
and their internal processes. However, similar to any other innovation,
AI also poses risks, and governments have the responsibility to maximize
the potentials and minimize the risks for their intended users. These
pros and cons influence the acceptance and adoption of AI in the public
sector. This study explored this angle in the context of the Philippine
government through a case study of one local government using the
combined constructs from TAM and UTAUT. A survey was conducted
to correlate the presumed drivers of attitude and behavioral intentions
toward use of AI in the local government unit. In-depth interviews were
thematically analyzed to determine the effect of the adoption factors
related to the attitude and intent to use AI, specified in TAM and UTAUT.
Although limited in making inferences about the causal relations

between the adoption factors, correlation results were consistent with
trends in prior TAM and UTAUT themed studies. Most acceptance and
adoption factors showed positive moderate to very strong relationships
with attitude and behavioral intention. On the other hand, anxiety only
showed significant negative moderate to very weak relationships with
AT.
Insights from the thematic analysis revealed themes that could

become the springboard for the national government to further improve
its existing national AI roadmap and develop appropriate policies and
Vol. 8: 33-65 (2021), ISSN 2718-9228
programs. One possible entry-point of the national government is

through the Department of Information and Communications Technology
(DICT), which has been a crucial actor in the enabling environment for
use of ICT technology in the country. The Department of the Interior
and Local Government (DILG), in its role to improve performance of
local governments in governance, administration, social and economic
development and environmental management, can also urge local
governments to institutionalize AI in relevant operational. An awareness
campaign along with a comprehensive upskilling program that includes
AI literacy for the citizenry and civil servants, can also be initiated by the
national government through DILG and DICT.
The Carmona LGU is a high-performing local government

with advanced ICT capability. Hence, the context of this research does
not encompass those of LGUs with a different level of capacity. Other
studies can validate the findings in other cases that could include but not
limited to (1) a high-performing and ICT-challenged local government,
(2) a low-performing and ICT-advanced local government, and (3) a low
performing and ICT-challenged local government. Such supporting case
studies can provide a richer understanding of the acceptance and adoption
perceptions on AI by other local governments. Moreover, expanded and
deepened research work can inform development of programmatic and
policy strategies for the scaling-out of AI in Philippine local governments
and beyond. In terms of research design, this study’s methods provided
valuable insights regarding the government acceptance and adoption
of AI, however, subsequent work should also accommodate the use of
analytical methods that would further test the causality of the acceptance
and adoption factors against the attitude and behavioral intention to use
AI. Further analysis may also check the mediating or moderating effect of
attitude to behavioral intention.
This paper acknowledges some of the criticisms of TAM and

UTAUT in past studies. Although UTAUT already addressed TAM’s
oversimplification of technology, both models still heavily rely on
an individual’s expectations and perceptions. It is also important to
consider the complexity of the process of technology adoption, from its
technical composition up to how it is integrated into the organization
(Ammenwerth, 2019; Shachak et al., 2019).
Future research can add the thematic analysis sub-themes

generated in this study to extend the TAM and UTAUT constructs of
adoption factors. For instance, the standard questionnaires only cover the
general usefulness, effectiveness, and productivity of AI to measure the
variable perceived usefulness. Future research may modify this by delving
deeper and adding statements discussing the concepts of performance

improvement, convenience, and decision-making. Other technology
acceptance models can also be explored with additional variables from
the field of motivational psychology.
Lastly, in this study, individual perceptions came from the key

decision-makers who reflect the local government’s vision. Though
findings may not be generalizable to all local government officials and
officers, they present valuable insights about the possible drivers of
acceptance and adoption of AI in government agencies. Future research
may re-validate the framework (Figure 4) to include other actors in
government, including street-level bureaucrats or citizen-clients, to
provide stronger empirical evidence on the acceptance and adoption of
AI by local governments.

REFERENCES
Alhashmi, S. F. S., Salloum, S. A., & Abdallah, S. (2019). Critical success

factors for implementing artificial intelligence (AI) projects in
Dubai government United Arab Emirates (UAE) health sector:
Applying the extended Technology Acceptance Model (TAM).
In A. Hassanien, K. Shaalan, & M. Tolba (Eds.), Proceedings of
the International Conference on Advanced Intelligent Systems
and Informatics 2019 (pp. 393–405). Springer. https://doi.
org/10.1007/978-3-030-31129-2_36
Al-Mushayt, O. S. (2019). Automating e-government services with
artificial intelligence. IEEE Access, 7, 146821–146829. https://doi.
org/10.1109/ACCESS.2019.2946204
Ammenwerth, E. (2019). Technology Acceptance Models in health
informatics: TAM and UTAUT. Applied Interdisciplinary Theory
in Health Informatics, 263, 64–71. https://doi.org/10.3233/
SHTI190111
Atwal, G., Bryson, D., & Williams, A. (2021). An exploratory study of the
adoption of artificial intelligence in Burgundy’s wine industry.
Strategic Change, 30(3), 299–306. https://doi.org/10.1002/jsc.2413
Baclig, C. E. (2021, June 21). Internet: Overcoming pandemic lockdowns.
Philippine Daily Inquirer. https://newsinfo.inquirer.net/1448837/
internet-overcoming-pandemic-lockdowns
Beresford Research. (2021). Age range by generation. https://www.
beresfordresearch.com/age-range-by-generation/
Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology.
Qualitative research in psychology, 3(2), 77–101. https://doi.
org/10.1191/1478088706qp063oa
Vol. 8: 33-65 (2021), ISSN 2718-9228
Buck, C., Doctor, E., Hennrich, J., Jöhnk, J., & Eymann, T. (2022). General
practitioners’ attitudes toward artificial intelligence–enabled
systems: Interview study. Journal of Medical Internet Research,
24(1), e28916. https://doi.org/10.2196/28916
Carrasco, M., Mills, S., Whybrew, A., & Jura, A. (2019, March 1). The
citizen’s perspective on the use of AI in government. Boston
Consulting Group. https://www.bcg.com/publications/2019/
citizen-perspective-use-artificial-intelligence-government-
digital-benchmarking
Creswell, J. W., & Creswell, J. D. (2018). Research design: Qualitative,
quantitative, and mixed methods approaches (5th ed.). SAGE
Publications, Inc.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user
acceptance of information technology. MIS Quarterly, 13(3), 319–
340. https://doi.org/10.2307/249008
Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of
computer technology: A comparison of two theoretical models.
Management Science, 35(8), 982–1003. https://doi.org/10.1287/
mnsc.35.8.982
Dickson, B. (2017, May 12). What is narrow, general and super artificial
intelligence. TechTalks. https://bdtechtalks.com/2017/05/12/
what-is-narrow-general-and-super-artificial-intelligence/
Department of Trade and Industry. (2021, May 5). DTI to launch National
Artificial Intelligence Roadmap. https://www.dti.gov.ph/archives/
national-artificial-intelligence-roadmap/
Dumlao-Abadilla, D. (2019, April 10). DTI: PH can become artificial
intelligence powerhouse. Philippine Daily Inquirer. https://
business.inquirer.net/268370/dti-ph-can-become-artificial-
intelligence-powerhouse
Fan, W., Liu, J., Zhu, S., & Pardalos, P. M. (2020). Investigating the
impacting factors for the healthcare professionals to adopt
artificial intelligence-based medical diagnosis support system
(AIMDSS). Annals of Operations Research, 294(1), 567–592.
https://doi.org/10.1007/s10479-018-2818-y
Girasa, R. (2020a). AI as a disruptive technology. In Artificial intelligence as
a disruptive Technology (pp. 3-21). Palgrave Macmillan.
Girasa, R. (2020b). Applications of AI and projections of AI impact. In
Artificial intelligence as a disruptive technology (pp. 23–67). Palgrave
Macmillan.
Gruzd, A., Staves, K., & Wilk, A. (2012). Connected scholars: Examining
the role of social media in research practices of faculty using the
UTAUT model. Computers in Human Behavior, 28(6), 2340–2350.
https://doi.org/10.1016/j.chb.2012.07.004
Hanif, Y., & Lallie, H. S. (2021). Security factors on the intention to use
mobile banking applications in the UK older generation (55+). A
mixed-method study using modified UTAUT and MTAM-with
perceived cyber security, risk, and trust. Technology in Society, 67,
101693. https://doi.org/10.1016/j.techsoc.2021.101693
Hasija, A., & Esper, T. L. (2022). In artificial intelligence (AI) we trust:
A qualitative investigation of AI technology acceptance. Journal
of Business Logistics. 43(3), 388–412. https://doi.org/10.1111/
jbl.12301
Hong, S. (2017). What are the areas of competence for central and
local governments? Accountability mechanisms in multi-level
governance. Journal of Public Administration Research and Theory,
27(1), 120–134. https://doi.org/10.1093/jopart/muw048
Hwang, S., & Nam, Y. (2020). The role of confidence in government in
acceptance intention towards artificial intelligence. Journal of
Digital Convergence, 18(8), 217–224. https://doi.org/10.14400/
JDC.2020.18.8.217
International Telecommunication Union. (2017a). Measuring the
Information Society Report 2017, Volume 1. https://www.itu.int/en/
ITU-D/Statistics/Documents/publications/misr2017/MISR2017_
Volume1.pdf
International Telecommunication Union. (2017b). Measuring the
Information Society Report 2017: Volume 2. ICT country profiles.
https://www.itu.int/en/ITU-D/Statistics/Documents/
publications/misr2017/MISR2017_Volume2.pdf
Jayaseelan, R., Kadeswaran, S., & Brindha, D. (2020). A qualitative
approach towards adoption of information and communication
technology by medical doctors applying UTAUT Model. Journal
of Xi’an University of Architecture & Technology, 12(3), 4689–4703.
http://ir.psgcas.ac.in/id/eprint/127
Kaya, T. (2019). Artificial intelligence driven e-government: The engage
model to improve e-decision making. In 19th European Conference
on Digital Government (pp. 43–50). Academic Conferences
International Limited. https://doi.org/10.34190/
Kessler, S. K. & Martin, M. (2017). How do potential users perceive the
adoption of new technologies within the field of Artificial Intelligence
and Internet-of-Things? [Master’s thesis, Lund University]. Lund
University Database. https://lup.lub.lu.se/luur
Kim, J. M. (2017). Study on intention and attitude of using artificial
intelligence technology in healthcare. Journal of Convergence for
Information Technology, 7(4), 53–60. https://doi.org/10.22156/
CS4SMB.2017.7.4.053
Vol. 8: 33-65 (2021), ISSN 2718-9228
Kim, J. W., Jo, H. I., & Lee, B. G. (2019). The study on the factors influencing
on the behavioral intention of chatbot service for the financial
sector: Focusing on the UTAUT Model. Journal of Digital Contents
Society, 20(1), 41–50. https://doi.org/10.9728/DCS.2019.20.1.41
Lee, A. S., & Baskerville, R. L. (2003). Generalizing generalizability in
information systems research. Information Systems Research,
14(3), 221–243. https://doi.org/10.1287/isre.14.3.221.16560
Lin, C., Hu, P. J. H., & Chen, H. (2004). Technology implementation
management in law enforcement: COPLINK system usability
and user acceptance evaluations. Social Science Computer Review,
22(1), 24–36. https://doi.org/10.1177%2F0894439303259881
Mehr, H. (2017). Artificial intelligence for citizen services and government.
Ash Center for Democratic Governance and Innovation,
Harvard Kennedy School. https://ash.harvard.edu/files/ash/files/
artificial_intelligence_for_citizen_services.pdf
Oxford Insights. (2020). Government AI Readiness Index 2020. https://
www.oxfordinsights.com/government-ai-readiness-index-2020
Plans and Initiatives of the Government on the Development in Artificial
Intelligence, S.R. 344, 17th Cong. (2017). http://legacy.senate.
gov.ph/lis/bill_res.aspx?congress=17&q=SRN-344
Proliferation of Artificial Intelligence-Synthesized Audiovisual Materials,
known as Deepfakes, S.R. 188, 18th Cong. (2019). http://legacy.
senate.gov.ph/lis/bill_res.aspx?congress=18&q=SRN-188
Rahi, S., Ghani, M., Alnaser, F., & Ngah, A. (2018). Investigating the role
of Unified Theory of Acceptance and use of Technology (UTAUT)
in Internet banking adoption context. Management Science Letters,
8(3), 173–186. http://dx.doi.org/10.5267/j.msl.2018.1.001
Shachak, A., Kuziemsky, C., & Petersen, C. (2019). Beyond TAM and
UTAUT: Future directions for HIT implementation research.
Journal of Biomedical Informatics, 100, 103315. https://doi.
org/10.1016/j.jbi.2019.103315
Sohn, K., & Kwon, O. (2020). Technology acceptance theories and factors
influencing artificial intelligence-based intelligent products.
Telematics and Informatics, 47, 101324. https://doi.org/10.1016/j.
tele.2019.101324
Stanford Human-Centered Artificial Intelligence. (2019). The Artificial
Intelligence Index 2019 Annual Report. https://hai.stanford.edu/
sites/default/files/ai_index_2019_report.pdf
United Nations. (2020). E-government survey 2020: Digital government in the
decade of action for sustainable development. United Nations Digital
Library. https://digitallibrary.un.org/record/3884686?ln=en
United Nations. (2022). E-government Survey 2022: The Future of Digital

Government. https://desapublications.un.org/sites/default/files/
publications/2022-09/Web%20version%20E-Government%20
2022.pdf.
Venkatesh, V. (2021). Adoption and use of AI tools: A research agenda
grounded in UTAUT. Annals of Operations Research, 308, 641–652.
https://doi.org/10.1007/s10479-020-03918-9
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User
acceptance of information technology: Toward a unified view.
MIS Quarterly, 27(3), 425–478. https://doi.org/10.2307/30036540
Williams, M. L., Saunderson, I.P., & Dhoest, A. (2021). Students’
perceptions of the adoption and use of social media in academic
libraries: A UTAUT study. Communication: South African Journal
of Communication Theory and Research, 47(1), 76–94. https://doi.or
g/10.1080/02500167.2021.1876123
Zhu, M., & Zhang, Y. (2021). Medical and public health instructors’
perceptions of online teaching: A qualitative study using the
Technology Acceptance Model 2. Education and Information
Technologies, 27, 2385–2405. https://doi.org/10.1007/s10639-021-
10681-2
Vol. 8: 33-65 (2021), ISSN 2718-9228
Appendix 1
Respondents’ demographic information
Characteristics Exploratory study In-depth interview

No. % No. %
(n=20) (n=6)a
Gender
Male 11 55.0 4 66.7
Female 9 45.0 2 33.3
Age
Millennial 3 15.0 2 33.3
Gen X 10 50.0 2 33.3
Boomer II 7 35.0 2 33.3
Cluster
Administrative/ 5 25.0 3 50.0
Legislative
Social 7 35.0 1 16.7
Infrastructure 4 20.0 1 16.7
Finance 4 20.0 1 16.7
Length of service in Carmona
1-10 years 4 20.0 1 16.7
11-20 years 5 25.0 2 33.3
21-above 11 55.0 3 50.0
Length of service in current position
1-10 years 10 50.0 3 50.0
11-20 years 4 20.0 1 16.7
21-above 6 30.0 2 33.3
five department heads and the municipal mayor

a
Appendix 2
Descriptive statistics per questionnaire statement of the exploratory survey
Var. Statement Obs. Mean Std. Dev. Min Max

PU PU1: My department and I find AI useful in our job. 20 5.55 1.145931 4 7
PU2: Using AI will enable me and my department to accomplish tasks quickly. 20 5.95 0.9445132 4 7
PU3: Using AI will increase my own and my department’s productivity. 20 5.70 1.128576 4 7
PU4: If I use AI, my department and I will enhance our job effectiveness. 20 5.80 1.056309 4 7
PE PE1: My department and my interaction with the AI will be clear and 20 5.30 1.174286 3 7
understandable.
PE2: It would be easy for me and my department to become skillful at using 20 5.55 1.190975 3 7
the AI.
PE3: My department and I will find the AI easy to use. 20 5.35 1.182103 3 7
PE4: Learning to operate the AI will be easy for me and my department. 20 5.50 1.192079 3 7
SI SI1: People who influence my behavior think I should use AI. 20 5.30 1.218282 3 7
SI2: People who are important to me think that I should use AI. 20 5.20 1.151658 3 7
SI3: The local government will be useful in the use of AI. 20 5.95 1.099043 4 7
SI4: In general, the local government will support the use of AI. 20 6.05 1.050063 4 7
PE: Perceived ease of use

PU: Perceived usefulness
SI: Social influence
Vol. 8: 33-65 (2021), ISSN 2718-9228
Appendix 2
Continued

FC FC1: My department and I have the resources necessary to use AI. 20 5.10 1.020836 3 7
FC2: My department and I have the knowledge necessary to use AI. 20 4.95 1.190975 2 7
FC3: AI is compatible with other systems we use in our department. 20 4.85 1.089423 2 6
FC4: A specific department or group will be available for assistance with AI 20 5.60 0.88258 4 7
difficulties.
SE SE1: My department and I could complete most tasks using AI if there will 20 4.50 1.60591 1 6
be no one around to tell us what to do as we use it.
SE2: My department and I could complete most tasks using AI if we could 20 5.55 1.099043 4 7
call someone for help if we get stuck.
SE3: My department and I could complete most tasks using AI if we just 20 5.70 1.080935 4 7
have the built-in help facility for assistance.
AX AX1: My department and I feel worried about using AI. 20 3.70 1.454575 1 6
AX2: It scares me and my department to think that we could lose a lot of 20 3.70 1.657519 1 6
information using AI if we click or move any wrong button or function.
AX3: My department and I hesitate to use AI for fear of making mistakes we 20 3.65 1.814416 1 6
cannot correct.
AX4: AI is somewhat intimidating to me and my department. 20 3.45 1.669384 1 6
AX: Anxiety
FC: Facilitating conditions
SE: Self-efficacy
Appendix 2
Continued

AT AT1: Using AI is a good idea. 20 6.05 1.050063 4 7
AT2: AI will make my department and my work more interesting. 20 6.10 1.020836 4 7
AT3: Working with AI seems fun. 20 5.60 1.046297 4 7
AT4: My department and I will like working with AI. 20 6.10 1.020836 4 7
BI BI1: My department and I intend to use AI in the next <n> months. 20 5.10 1.333772 2 7
BI2: My department and I would use AI in the next <n> months. 20 5.05 1.394538 2 7
BI3: My department and I plan to use AI in the next <n> months. 20 5.10 1.252366 2 7
AT: Attitude toward using AI

BI: Behavioral intention to use AI

798-Article Text-2442-1-10-20230109

Uploaded by

Copyright:

Available Formats

798-Article Text-2442-1-10-20230109

Uploaded by

Document Information

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

798-Article Text-2442-1-10-20230109

Uploaded by

Copyright:

Available Formats

Journal of Public Affairs and Development

Vol. 8: 33-65 (2021), ISSN 2718-9228

Adoption of Artificial Intelligence (AI)

ABSTRACT. Emerging technologies like artificial intelligence (AI) have

Keywords: artificial intelligence, local government, TAM, UTAUT,

Technology is ever evolving, and these changes cause some

Various governments have started integrating AI into their

This paper addresses the research question: how do Philippine

AI Applications and Risks in Digital Government

AI is one of the most common disruptive technologies, which is

Be data-prepared Mitigate ethical Augment

Fast-forward to the 21st century, AI is becoming more relevant

Dickson (2017) categorized AI into three types. First, artificial

Similar to other e-government efforts, AI has its fair share of

Current State of AI Applications and Digital Government

Based on the 2022 E-government Survey of the United Nations,

However, the Philippine government still needs to focus on

Nevertheless, some improvements in the Philippine ICT

In terms of AI adoption, the Philippines is tagged as one of the

Unlike countries with established AI strategies and policies,

an interview that the country seeks to be known as a“global AI hub”. He

As early as 2017, Filipino policymakers also showed interest

Model and Theory on Technology Acceptance

AI alongside other emerging and disruptive technologies are

Besides these studies, there is still an existing research gap on

Technology Acceptance Model (TAM). Davis (1989) argued

Unified Theory of Acceptance and Use of Technology

Another key factor that they added is social influence, defined

Guided by the TAM and UTAUT, a case study was conducted to

The case study was conducted in the Municipality of Carmona

Mixed Methods Approach

This study used a mixed methods approach to capture the

The authors deployed a self-administered survey to all 20

Thematic analysis was conducted to explain and contextualize

In total, 21 open-ended questions around the concepts

RESULTS AND DISCUSSION

Descriptive Statistics and Correlation Analysis

Table 1 shows the overall mean scores of all AI acceptance and

Variable Overall mean score

Social influence 5.63

Meanwhile, facilitating conditions received an overall mean of

PU 0.8373* 0.7476* 1.0000

PE 0.8501* 0.7722* 0.8757* 1.0000

SI 0.6930* 0.8136* 0.8170* 0.7586* 1.0000

FC 0.6895* 0.5007** 0.5905* 0.6453* 0.5484** 1.0000

SE 0.5647* 0.4073*** 0.5979* 0.6990* 0.5202** 0.5112** 1.0000

AX -0.3002 -0.5490** -0.5445** -0.5360** -0.6275* -0.1653 -0.3617 1.0000

Notes: *p<0.01, **p<0.05, ***p<0.10 (2-tailed)

PE: Perceived ease of use AT: Attitude toward using AI

intention (r = 0.7224), which can be construed as a positive attitude of

Meanwhile, for the correlation of attitude toward using AI, social

Several AI acceptance and adoption factors have also shown very

be indirect drivers of behavioral intention. Furthermore, all the observed

To further explain the correlation findings, the in-depth

Previous Use of AI. The respondents had prior experience in

Perceived Usefulness. Most of the themes that emerged from

The municipal administrator and IT department head both

Perceived Ease of Use. Majority of the respondents agreed that

However, most of the respondents agreed that the ease of use

The IT department head highlighted that most of the time, people

SE 0.5647* 0.4073*** 0.5979* 0.6990* 0.5202 0.5112 1.0000

AX -0.3002 -0.5490 -0.5445 -0.5360** -0.6275* -0.1653 -0.3617 1.0000

Notes: p<0.01, p<0.05, p<0.10 (2-tailed)