1 s2.0 S004016252400043X Main

Technological Forecasting & Social Change 201 (2024) 123247
Contents lists available at ScienceDirect
Technological Forecasting & Social Change

journal homepage: www.elsevier.com/locate/techfore
Factors affecting performance expectancy and intentions to use ChatGPT:

Using SmartPLS to advance an information technology
acceptance framework
Mark Anthony Camilleri *
Department of Corporate Communication, Faculty of Media and Knowledge Sciences, University of Malta, Msida MSD2080, Malta
Northwestern University, Medill School, 1845 Sheridan Rd, Evanston, IL 60208, United States
The Business School, Buccleuch Place, EH8 9JS Edinburgh, Mid-Lothian, Scotland, United Kingdom.
A R T I C L E I N F O A B S T R A C T
Keywords: Few studies have explored the use of artificial intelligence-enabled (AI-enabled) large language models (LLMs).
Unified Theory of Acceptance and use of This research addresses this knowledge gap. It investigates perceptions and intentional behaviors to utilize AI
Technology dialogue systems like Chat Generative Pre-Trained Transformer (ChatGPT). A survey questionnaire comprising
Information Adoption Model
measures from key information technology adoption models, was used to capture quantitative data from a
Chat Generative Pre-Trained Transformer
ChatGPT
sample of 654 respondents. A partial least squares (PLS) approach assesses the constructs’ reliabilities and
AI Chatbot validities. It also identifies the relative strength and significance of the causal paths in the proposed research
Natural language generation model. The findings from SmartPLS4 report that there are highly significant effects in this empirical investigation
particularly between source trustworthiness and performance expectancy from AI chatbots, as well as between
perceived interactivity and intentions to use this algorithm, among others. In conclusion, this contribution puts
forward a robust information technology acceptance framework that clearly evidences the factors that entice
online users to habitually engage with text-generating AI chatbot technologies. It implies that although they may
be considered as useful interactive systems for content creators, there is scope to continue improving the quality
of their responses (in terms of their accuracy and timeliness) to reduce misinformation, social biases, halluci
nations and adversarial prompts.
1. Introduction created by humans (such content is accessed through the Internet). The
responses it provides appear to be as human-like as possible (Jiang et al.,
Artificial intelligence (AI) chatbots utilize algorithms that are trained 2023).
to process and analyze vast amounts of data by using techniques ranging GPT-3.5’s database was last updated in September 2021. However,
from rule-based approaches to statistical models and deep learning, to GPT-4.0 version comes with a paid plan that is more creative than GPT-
generate natural text, to respond to online users, based on the input they 3.5, could accept images as inputs, can generate captions, classifications
received (OECD, 2023). For instance, Open AI’s Chat Generative Pre- and analyses (Qureshi et al., 2023). Its developers assert that GPT-4.0
Trained Transformer (ChatGPT) is one of the most popular AI- can create better content including extended conversations, as well as
powered chatbots. The company claims that ChatGPT “is designed to document search and analysis (Takefuji, 2023). Recently, its proponents
assist with a wide range of tasks, from answering questions to generating noted that ChatGPT can be utilized for academic purposes, including
text in various styles and formats” (OpenAI, 2023a). OpenAI clarifies research. It can extract and paraphrase information, translate text, grade
that its GPT-3.5, is a free-to-use language model that was optimized for tests, and/or it may be used for conversation purposes (MIT, 2023).
dialogue by using Reinforcement Learning with Human Feedback Various stakeholders in education noted that this LLM tool may be able
(RLHF) – a method that relies on human demonstrations and preference to provide quick and easy answers to questions.
comparisons to guide the model toward desired behaviors. Its models However, earlier this year, several higher educational institutions
are trained on vast amounts of data including conversations that were issued statements that warned students against using ChatGPT for
* Department of Corporate Communication, Faculty of Media and Knowledge Sciences, University of Malta, Msida MSD2080, Malta.
E-mail address: [email protected].
https://doi.org/10.1016/j.techfore.2024.123247
Received 16 August 2023; Received in revised form 9 January 2024; Accepted 22 January 2024
Available online 6 February 2024
0040-1625/© 2024 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
M.A. Camilleri Technological Forecasting & Social Change 201 (2024) 123247
academic purposes. In a similar vein, a number of schools banned Planned Behavior (TPB) (Ajzen, 1991), the Technology Acceptance
ChatGPT from their networks and devices (Rudolph et al., 2023). Model (TAM) (Davis, 1989; Davis et al., 1989), TAM2 (Venkatesh and
Evidently, policy makers were concerned that this text generating AI Davis, 2000), TAM3 (Venkatesh and Bala, 2008), UTAUT (Venkatesh
system could disseminate misinformation and even promote plagiarism. et al., 2003) or UTAUT2 (Venkatesh et al., 2012). Few scholars have
Some commentators argue that it can affect the students’ critical- integrated constructs like UTAUT/UTAUT2’s performance expectancy,
thinking and problem-solving abilities. Such skill sets are essential as effort expectancy, social influences and intentions to use technologies
pects for their academic and lifelong successes (Liebrenz et al., 2023; with information quality and source trust measures from the Elaboration
Thorp, 2023). Nevertheless, a number of jurisdictions are reversing their Likelihood Model (ELM) and IAM. Currently, there is still limited
decisions that impede students from using this technology (Reuters, research that incorporates a perceived interactivity factor within infor
2023). In many cases, educational leaders are realizing that their stu mation technology frameworks. Therefore, this contribution addresses
dents could benefit from this innovation, if they are properly taught how this deficit in academic knowledge.
to adopt it as a tool for their learning journey. Notwithstanding, for the time being, there is still scant research that
Academic colleagues are increasingly raising awareness on different is focused on AI-powered LLM, like ChatGPT, that are capable of
uses of AI dialogue systems like service chatbots and/or virtual assis generating human-like text that is based on previous contexts and drawn
tants (Baabdullah et al., 2022; Balakrishnan et al., 2022; Brachten et al., from past conversations. This timely study raises awareness on the in
2021; Hari et al., 2022; Li et al., 2021; Lou et al., 2022; Malodia et al., dividuals’ perceptions about the utilitarian value of such interactive
2021; Sharma et al., 2022). Some of them are evaluating their strengths technologies, in an academic (higher educational) context. It clearly
and weaknesses, including of OpenAI’s ChatGPT (Farrokhnia et al., identifies the factors that are influencing the individuals’ intentions to
2023; Kasneci et al., 2023). Very often, they argue that there may be continue using them, in the future.
instances where the chatbots’ prompts are not completely accurate and/ The following section introduces the readers of this article to rele
or may not fully address the questions that are asked to them (Gill et al., vant theoretical underpinnings that lead to the formulation of hypoth
2024). This may be due to different reasons. For example, GPT-3.5’s eses. It presents a graphical illustration of the proposed research model.
responses are based on the data that were uploaded before a knowledge Then, the methodology sheds light on the measures that were used to
cut-off date (i.e. September 2021). This can have a negative effect on the capture the quantitative data. It also provides information on the survey
quality of its replies, as the algorithm is not up to date with the latest administration and describes the profile of the research participants.
developments. Although, at the moment, there is a knowledge gap and a Afterwards, the data analysis features the descriptive statistics as well as
few grey areas on the use of AI chatbots that use natural language the findings from partial least squares (PLS) algorithm and from the
processing to create humanlike conversational dialogue, currently, there Bootstrapping procedure. This section reveals the results of the hy
are still a few contributions that have critically evaluated their pros and pothesis testing. In conclusion, this contribution puts forward its theo
cons, and even less studies have investigated the factors affecting the retical and managerial implications. It identifies the limitations of this
individuals’ engagement levels with ChatGPT. study and presents future research avenues.
This empirical research builds on theoretical underpinnings related
to information technology adoption in order to examine the online 2. Conceptual framework
users’ perceptions and intentions to use AI Chatbots. Specifically, it in
tegrates a perceived interactivity construct (Baabdullah et al., 2022; 2.1. Intentions to use information technologies
McMillan and Hwang, 2002) with information quality and source
trustworthiness measures (Leong et al., 2021; Sussman and Siegal, 2003) There are various theoretical underpinnings that are intended to
from the Information Adoption Model (IAM) with performance expec explain the factors affecting the users’ intentions to utilize a wide array
tancy, effort expectancy and social influences constructs (Venkatesh of technologies. For instance, the Theory of Planned Behavior (TPB) as
et al., 2003; Venkatesh et al., 2012) from the Unified Theory of Accep well as its predecessor, the Theory of Reasoning Action (TRA) postulate
tance and Use of Technology (UTAUT1/UTAUT2) to determine which that the persons’ attitudes and beliefs would anticipate their behaviors
factors are influencing the individuals’ intentions to use AI text gener and actions. The former comprises three antecedents, namely, attitudes,
ation systems like ChatGPT. This study’s focused research questions are: subjective norms, and perceived behavioral control. Ajzen (1991)
argued that these constructs affect the individuals’ behavioral in
RQ1. How and to what extent are information quality and source
tentions. Previously, Fishbein and Ajzen’s (1975) TRA posited that at
trustworthiness influencing the online users’ performance expectancy
titudes and subjective norms anticipate the persons’ intentions to
from ChatGPT?
engage in certain activities. TRA is very similar to TPB, albeit it does not
RQ2. How and to what extent are their perceptions about ChatGPT’s include a perceived behavior control construct.
interactivity, performance expectancy, effort expectancy, as well as their The behavioral intention can be defined as an individual’s readiness
social influences affecting their intentions to continue using their large to perform given behaviors. Ajzen (1991) contended that the in
language models? dividuals’ intentions are assumed to capture the motivational factors
that influence behavior. Essentially, they represent the extent to which
RQ3. How and to what degree is the performance expectancy
people are willing to engage in certain activities, or to make an effort to
construct mediating effort expectancy – intentions to use these inter
perform specific behaviors. This construct has been widely utilized to
active AI technologies?
explore human actions in various contexts including in psychology,
This study hypothesizes that information quality and source trust cognitive studies, marketing, information management, technology
worthiness are significant antecedents of performance expectancy. It adoption, et cetera.
presumes that this latter construct, together with effort expectancy, Several researchers relied on the behavioral intentions measure to
social influences as well as perceived interactivity affect the online examine the persons’ dispositions to adopt information systems.
users’ acceptance and usage of generative pre-trained AI chatbots like Frequently, it has been featured as an endogenous factor in various
GPT-3.5 or GPT-4. studies, including in the Technology Acceptance Model (TAM) (Davis,
Many academic researchers sought to explore the individuals’ 1989; Davis et al., 1989; Venkatesh and Davis, 2000). Like TPB and TRA,
behavioral intentions to use a wide array of technologies (Alalwan, TAM presumes that the persons’ attitudes about information commu
2020; Alam et al., 2020; Al-Saedi et al., 2020; Raza et al., 2021; Tam nication technologies (ICT) are a precursor for their behavioral in
et al., 2020). Very often, they utilized measures from the Theory of tentions to use them. In addition, it posits that the individuals’
Reasoned Action (TRA) (Fishbein and Ajzen, 1975), the Theory of perceptions about the usefulness as well as the ease-of-use of ICT can
2
influence their attitudes and intentions to use them. usefulness and perceived ease-of-use. Yet, the behavioral intentions to
These theoretical models and their key factors have continuously use technology devices and their systems were frequently hypothesized
been validated in different studies (Driediger and Bhatiasevi, 2019; Ho to be an endogenous construct. Table 1 features a summary of the most
et al., 2020; Kamble et al., 2019; Sohn and Kwon, 2020). Very often, popular theoretical frameworks that sought to identify the antecedents
they were expanded, as many researchers integrated new constructs in and the extent to which they may affect the individuals’ intentions to use
extant models. For example, TAM2 (Venkatesh and Davis, 2000) and information technologies.
TAM3 (Venkatesh and Bala, 2008) featured additional factors to TAM.
The authors sought to investigate the antecedents of perceived
Table 1
A non-exhaustive list of theoretical frameworks focused on (information) technology adoption behaviors.
Theory Developed by Measures
Electronic Service Quality Parasuraman Ease of use Website design Reliability System Privacy
(eSERVQUAL) et al. (2005) availability
Responsiveness Empathy Consumer Trust Satisfaction
experience
Electronic retail quality Wolfinbarger Consumer Website design Privacy/security Customer service Satisfaction
(eTailQ) and Gilly (2003) fulfillment
Information Adoption Sussman and Argument quality Source Information Information
Model (IAM) Siegal (2003) credibility usefulness adoption
Information technology Camilleri et al. Information Source Functionality Perceived Intention to use
adoption model (2023) quality credibility usefulness the technology
(InfTAM)
Innovation diffusion Moore and Relative advantage Ease of use Image Voluntarism
theory (IDT/DOI) Benbasat (1991)
Compatibility Visibility Result Adoption
demonstrability
Interactive technology Camilleri and Information Source Interactive Perceived Intentions to use
adoption model (ITAM) Kozak (2023) quality credibility engagement usefulness the technology
Model of PC utilization Thompson et al. Social factors Affect toward PC Complexity of PC Job fit with PC
(MPCU) (1991) use
Long term Facilitating PC utilization
consequences of conditions for PC
PC use use
Motivation model (MM) Davis et al. Task importance Perceived ease of Perceived Output quality
(1992) use usefulness
Perceived Intentions to use Usage behavior
enjoyment the technology
Net quality (NETQual) Bressolles et al. Ease of use Information Design Reliability Security/privacy Interactivity/
(2014) personalization
Site quality (SITEQUAL) Yoo and Donthu Ease of use Aesthetic design Processing speed Security
(2001)
Social cognitive theory Compeau and Prior performance Computer self- Behavior Outcome Performance
(SCT) Higgins (1995) efficacy modelling expectations
Synchronous technology Camilleri and Functionality Perceived Facilitating Intentions to use
adoption model (STAM) Camilleri (2022) interactivity conditions the technology
Technology acceptance Davis et al. Perceived ease of Perceived Attitudes Intentions to use Actual behaviors
model (TAM) (1989) use usefulness the technology
Technology acceptance Venkatesh and Subjective norms Image Job relevance Output quality Result
model (enhanced Davis (2000) demonstrability
version TAM2) Experience Voluntariness Perceived Perceived ease of Intentions to use Actual behaviors
usefulness use the technology
Technology acceptance Venkatesh and Subjective norms Image Job relevance Output quality Computer Result
model (enhanced Bala (2008) playfulness demonstrability
version TAM3) Result Computer self- Perception of Computer Perceived ease of Perceived
demonstrability efficiency external control anxiety use usefulness
Experience Voluntariness Intentions to use Actual behaviors
the technology
Theory of planned Ajzen (1991) Attitudes Subjective norms Perceived Intentions to use Actual behaviors
behavior (TPB) behavioral control the technology
Theory of reasoned action Fishbein and Attitudes Subjective norms Intentions to use Actual behaviors
(TRA) Ajzen (1975) the technology
Transaction process-based Bauer et al. Responsiveness Reliability Process Functionality/ Enjoyment
approaches for (2006) design
capturing service
quality (eTransQual)
Unified Theory of Venkatesh et al. Performance Effort Social influence Facilitating Gender
Acceptance and Use of (2003) expectancy expectancy conditions
Technology (UTAUT1) Age Experience Voluntariness in Intentions to use Actual behaviors
use the technology
Unified Theory of Venkatesh et al. Performance Effort Social influence Facilitating Hedonic Price value
Acceptance and Use of (2012) expectancy expectancy conditions motivations
Technology (Enhanced Habit Age Gender Experience Intention to use Actual behaviors
version - UTAUT2) the technology
Uses and gratifications Camilleri and Perceived ease of Perceived Ritualized use Instrumental use Intention to use
theory (U&G) Falzon (2021) use usefulness the technology
3
2.2. Performance expectancy mediates effort expectancy - intentions to use this information
technology.
In this case, this research investigates the extent to which users
H3. The online users’ effort expectancy of ChatGPT significantly af
accept and use specific technologies. It builds on Venkatesh et al.’s
fects the performance expectancy of this information technology.
(2003) UTAUT. These authors referred to performance expectancy when
they advanced their theoretical model. They indicated that their “new”
2.4. Information quality
construct was similar to Davis’ (1989) perceived usefulness as well as to
other factors, including to extrinsic motivation from Davis et al.’s (1992)
Generally, the technology adoption models including the extended
Motivation Model, job-fit from Thompson et al.’s (1991) Model of PC
TAM and UTAUT frameworks, among others, are meant to measure the
Utilization, relative advantage from Moore and Benbasat’s (1991)
persons’ dispositions to avail themselves of information technologies,
Innovation Diffusion Theory, and to outcome expectations, that is drawn
and to identify which factors and to what extent they affect their in
from Compeau and Higgin’s (1995) Social Cognitive Theory. Venkatesh
tentions to engage with. Them. The variants of the Elaboration Likeli
et al. (2003) noted that these factors intended to measure how the use or
hood Model (ELM) and/or Information Adoption Model (IAM) are
capabilities of a system were instrumental in improving the individuals’
intended to clarify whether central and/or peripheral aspects of
job performance, by accomplishing tasks, by enhancing productivity,
communication could influence the individuals’ attitudes about
and/or by increasing their chances of getting a pay rise.
persuasive content (Cacioppo and Petty, 1981), and/or their perceptions
In the authors’ own words, they defined performance expectancy as
about the usefulness of information (Sussman and Siegal, 2003).
“the degree to which an individual believes that using the system will
Arguably, in the latter case, the communicated content can influence
help him or her attain gains in job performance”. When Venkatesh et al.
the persons’ perceptions in different ways, as they may not always be
(2003) put forward their UTAUT, they theorized that the individuals’
willing to evaluate the quality of the messages they receive (Sussman
performance expectancy from technological systems significantly affects
and Siegal, 2003). This argumentation is synonymous with ELM’s cen
their behavioral intentions to use them. Since then, a number of re
tral route, as recipients of information may opt to carefully assess the
searchers have reported similar results, even though they examined and
content that is presented them. ELM suggests that the quality of elabo
validated this causal path for a wide array of technologies, in various
rated information can influence the individuals’ attitudes toward the
contexts (Alalwan, 2020; Alam et al., 2020; Al-Saedi et al., 2020; Raza
message. Cacioppo and Petty (1981) contended that people reflect on
et al., 2021; Tam et al., 2020). In a similar vein, this research presumes
the content of the message that is conveyed to them. These authors
that:
maintained that the recipients of information arrive at a reasoned atti
H1. The online users’ performance expectancy of ChatGPT signifi tude that is supported by the argument quality. Cacioppo and Petty
cantly affects their intentions to use this information technology. (1981) among others (E.g. Cheung et al., 2008; Erkan and Evans, 2018;
Sussman and Siegal, 2003) clearly distinguished between central and
peripheral factors that can influence how individuals elaborate on a
2.3. Effort expectancy
persuasive message.
Sussman and Siegal (2003) have built their IAM theoretical un
Venkatesh et al. (2003) maintained that their effort expectancy
derpinnings on the foundations of ELM. Their research model suggested
construct is particularly significant during the earlier stages when in
that information quality is a significant antecedent of information use
dividuals start acquainting themselves with the technology, in manda
fulness. Over the years, a number of researchers have validated this
tory, as well as in voluntary contexts. They postulated that the
causal path in different contexts (Jin et al., 2009; Leong et al., 2021;
individuals’ ease of use of technology becomes an insignificant factor
Peng et al., 2016). Erkan and Evans (2018) reported that information
over periods of habitual behaviors, after extended usage. These authors
quality significantly affects the usefulness of information technologies
noted that UTAUT’s effort expectancy construct is synonymous with
like shopping websites that feature consumer recommendations. Their
TAM’s and TAM2’s perceived ease-of-use as well as with the Model of PC
study indicated that their respondents appreciated their utilitarian value
Utilization’s complexity construct and with ease-of-use from Innovation
as they were willing to rely on their user generated content. Similarly,
Diffusion Theory. Venkatesh et al. (2003) noted that the definitions of
Camilleri and Filieri (2023) reported that information quality signifi
these constructs as well as their measuring items were very similar.
cantly affects information usefulness. The latter construct is synonymous
Whilst Davis (1989) clarified that perceived ease-of-use refers to “the
with Davis (1989) perceived usefulness factor. Hence, it is related to
degree to which a person believes that using a particular system would
Venkatesh et al.’s (2012) performance expectancy. This research
be free of effort”, Venkatesh et al. (2003) suggested that; “effort ex
hypothesizes:
pectancy is the degree of ease associated with the use of the system”.
In the main, information systems researchers argue that individuals H4. The online users’ perceptions about the information quality of
would be intrigued to use clear and understandable technologies that are ChatGPT’s generated responses significantly affect their performance
uncomplicated and easy to use (Thompson et al., 1991; Venkatesh et al., expectancy of this information technology.
2012). Most academic authors hypothesize that the persons’ effort ex
pectancy from certain technologies, in terms of their systems’ ease-of- 2.5. Source trustworthiness
use, would increase their likelihood to continue using them in the
future (Abbad, 2021; Beh et al., 2021; Queiroz et al., 2021). Previously, At times, individuals may ignore the quality of the messages they
Davis’s (1989) as well as Davis et al.’s (1989) TAM indicated that receive. They may do so as they hold preconceived perceptions about the
perceived ease of use also had a significant effect on perceived useful source of the message. This reasoning is related to ELM’s theoretical
ness. However, unlike TAM studies (Davis, 1989; Venkatesh and Davis, underpinnings. Its proponents argue that according to the peripheral
2000; Venkatesh and Bala, 2008), few UTAUT/UTAUT2 studies have route to communication, there may be different reasons that could affect
examined the effects of effort expectancy on performance expectancy of the receivers’ willingness to accept and process the messages that are
information technologies (Camilleri and Kozak, 2022). This argumen conveyed to them (Cacioppo and Petty, 1981; Shi et al., 2018). They
tation leads to the following hypotheses: may not be interested in examining the content they receive. Hence,
they will probably rely on subjective cues and on their general impres
H2. The online users’ effort expectancy of ChatGPT significantly af
sions about the source (Ferguson and Mohan, 2020). ELM commentators
fects their intentions to use this information technology.
generalize that individuals could be influenced by heuristics (mental
H2a. The online users’ performance expectancy of ChatGPT partially shortcuts), particularly if they are not motivated/interested in the
4
messages that are transmitted to them (Bingham et al., 2019; Rohde and conduct specific activities including adopting new technologies (Kamal
Mau, 2021). Alternatively, they may be distracted from scrutinizing et al., 2020; Patil et al., 2020; Raza et al., 2021; Zhao and Bacao, 2020).
their content, and/or do not possess the cognitive abilities and/or may Hence, this study hypothesizes:
not have the time/opportunity to do so (Hoeken et al., 2020). In this
H6. Social influences would significantly affect the online users’ in
case, the targeted audience may rely on the sources’ trustworthiness,
tentions to use ChatGPT.
rather than on the quality of their arguments.
Various researchers found that peripheral cues in websites and/or
social media can have an impact on the individuals’ perceptions, atti 2.7. Perceived interactivity
tudes, as well as on intentional behaviors, when they are unable or
unwilling to elaborate on the message’s content (John and De’Villiers, The users of information technologies would usually appraise the
2020; Winter, 2020). Very often, they indicated that the persons’ atti systems’ attributes and features with colleagues (as well as with family
tudes toward communications are influenced by the source’s attrac and friends), particularly if they found them to be useful, functional
tiveness, likeability, and credibility in terms of their trustworthiness and and/or if they exceeded their expectations in terms of their interactivity
expertise (Cacioppo and Petty, 1981). Sussman and Siegal (2003) noted aspects. Most world-wide web technologies including blogs, social
that source credibility positively influences the users’ acceptance of media, review sites, web chatbots, virtual assistants, and the like, can be
knowledge-based systems’ recommendations. They went on to suggest considered to be interactive, as they involve two-way communications.
that credible sources are persuasive and can play an important role in Yet, the notion of interactivity is often misunderstood, unexplained or
informational influence. Erkan and Evans (2016) reported that the in underdefined. Relevant academic literature suggests that: (i) the direc
dividuals would rely on the information obtained from social media if tion of communication, (ii) user control, and (iii) time are three over
they perceive that the source of the content is credible and dependable. lapping constructs that can describe the interactivity features of various
A number of empirical studies confirmed that source trustworthiness (or technologies (McMillan and Hwang, 2002). Firstly, the direction of
source credibility) is an antecedent of the individuals’ perceptions about communication is related to the degree of responsiveness and to the
the usefulness of information (Camilleri and Filieri, 2023; Kang and exchange of information (Bauer et al., 2006; Parasuraman et al., 2005).
Namkung, 2019; Onofrei et al., 2022). In a similar vein, this research Secondly, user control is associated with functions such as the extent of
presumes that this construct can be a significant precursor of perfor concurrent participation and to interpersonal, online engagement.
mance expectancy. Thus, this study hypothesizes: Thirdly, the concept of time is linked to the timeliness of immediate
feedback.
H5. The online users’ perceptions about the source trustworthiness of
Web content can be accessed through functional and easy-to-use
ChatGPT’s generated responses significantly affect their performance
navigational tools like digital devices and mobile applications (Camil
expectancies of this information technology.
leri and Camilleri, 2022; Kaya et al., 2019; Molinillo et al., 2020). These
systems are also meant to facilitate human-to-human, human-to-com
2.6. Social influences puter as well as computer-to-human interactions. Online users can
create and share their vocal, verbal and visual content with others. They
The individuals’ acceptance and usage of information technologies present themselves through their interactive content. This argumenta
can be triggered by social influences. This issue is conspicuous with the tion is synonymous with the social exchange theory that suggests that
subjective norm dimension that was discussed in Fishbein and Ajzen’s there is scope for individuals to reciprocate with others as they can
(1975) TRA as well as in Ajzen’s (1991). The subjective norm construct obtain informational as well as emotional values from interactions
raises awareness about possible peer pressures and general beliefs on (Cortez and Johnston, 2020; Luqman et al., 2023). Several researchers
different issues, from family, friends, work colleagues and from other sought to explore the individuals’ perceptions and experiences with
members in society, on the individuals’ intentional behaviors and ac interactive media. Song and Zinkhan (2008) posited that the presence or
tivities including on their engagement with technology. absence of particular design features (e.g. choice of background colors,
Venkatesh and Davis (2000) held that there are three interrelated search options, clickable areas, feedback mechanisms, et cetera) can
social forces that can impinge on the users’ adoption or rejection of a determine the interactivity levels of certain technologies.
new system, including subjective norm, voluntariness, and image. They A number of academic commentators made reference to a perceived
explained that the rationale behind the direct effect of subjective norm interactivity construct to examine the individuals’ self-presentation
on intention is that individuals tend to engage in behaviors, even if they (online), their content contribution as well as their exchange of sup
do not like them or their consequences. Yet, they may decide to engage port with other users (Zhang et al., 2014). Zhao and Lu (2012) distin
in certain activities if the persons around them think they should. These guished between two dimensions: user-to-user interactivity (i.e.
authors distinguished between mandatory and voluntary usage contexts. interpersonal online engagement) and user-to-system (human-machine
Subsequently, a number of academic colleagues specifically referred interactivity).
to a social influences construct (Camilleri and Kozak, 2022; Venkatesh Recently, various researchers have even investigated whether the
et al., 2003; Venkatesh et al., 2012) in UTAUT/UTAUT2, or to social persons’ perceptions about interactive artificial intelligence systems like
factors (Thompson et al., 1991) in the Model of PC Utilization, or to chatbots or virtual assistants for customer services purposes (Peltier
image (Moore and Benbasat, 1991) in Innovation Diffusion Theory. Very et al., 2023). Very often they reported that they were satisfied with the
often, the researchers who used one of these constructs, indicated that dialogue systems’ interactivity features, as they have significantly
the individuals are expected to comply with certain norms and values in affected their intentions to continue using them in the future (Baab
society, particularly in mandatory environments. Such social influences dullah et al., 2022; Lou et al., 2022). Similarly, this research presumes
can also influence their perceptions about the acceptance of information that:
technologies that can be utilized for different purposes. They could even
H7. The online users’ perceptions about the interactivity of ChatGPT
induce their adoption.
significantly affect their intentions to use this information technology.
Venkatesh et al. (2003) suggested that this construct is very impor
tant in the early stages of the individuals’ experiences with technologies Fig. 1 features the conceptual framework that investigates informa
(this reasoning is also congruent with the perceived ease-of-use or effort tion technology adoption factors. It represents a visual illustration of the
expectancy constructs), as they will not have to follow their peers’ hypotheses of this study. In sum, this empirical research presumes that
recommendations once they become habituated with them. Several information quality and source trustworthiness (from Information
studies confirmed that in many cases individuals are urged by others to Adoption Model) precede performance expectancy. The latter construct
5
Fig. 1. Information technology acceptance framework.
together with effort expectancy, social influences (from Unified Theory group of academic colleagues.
of Acceptance and Use of Technology) as well as the perceived inter The research participants disclosed their gender as well as their age
activity construct, are significant antecedents of the individuals’ in by choosing one of five age groups in the last part of the survey. They
tentions to use ChatGPT. also indicated the highest qualification that they attained at the time of
this study.
3. Methodology
3.3. The demographic profile of the respondents
3.1. The survey administration
After a few weeks, there were six hundred fifty-four (n = 654) re
Primary data were collected through an online survey questionnaire spondents who confirmed (through a filter question) that they have used
that was disseminated via an email, among all members of staff as well ChatGPT. The frequency table reported 292 males, 338 males and 22
as students who were enrolled in full time and part time courses in a participants who opted not to indicate their gender. The research par
Southern European university, during the second semester of ticipants were categorized into 5 age groups (18–28; 29–39; 40–50;
2022–2023. There were >13,200 research participants who were tar 51–61; Over 62). The majority of them were between 18 and 28 years (n
geted for this research about the use of ChatGPT. They were in a position = 318). The second largest group involved middle-aged individuals who
to complete the questionnaire in a few minutes. were between 40 and 50 years (n = 128). Most of the respondents re
This empirical study complied with the research ethic policies of the ported that they had completed an undergraduate level of education, as
higher educational institution as well with the EU’s (2016) general data indicated in Table 3.
protection regulations (GDPR). There was no way that the survey re
spondents’ identity could be revealed, as only aggregate data was 4. Data analysis
required for this quantitative research.
4.1. The descriptive statistics
3.2. The survey instrument
The findings reported that, in the main, the research participants
The respondents were instructed to answer all survey questions that were agreeing with the statements that were presented to them in the
were presented to them about information quality, source trustworthi survey questionnaire. The mean values were mostly above 3. Whilst EE1
ness, performance expectancy, effort expectancy, social influences, (M = 4.239) and EE2 (M = 4.18) were the highest mean scores, PI2 (M
perceived interactivity and on their behavioral intentions to continue = 2.908) and IQ2 (M = 2.911) reported the lowest means. The SD values
using this technology (otherwise, they could not submit the question were relatively low as the highest variance figure was 1.216 (for SI1).
naire). Table 2 features the list of measures as well as their corre
sponding items that were utilized in this study. It also provides a 4.2. Results from PLS-SEM algorithm
definition of the constructs used in the proposed information technology
acceptance framework. A collinearity assessment revealed that there was no evidence of
The research participants were expected to clearly indicate the common method bias in this study. Table 4 illustrates the results of the
extent of their agreement with the survey’s measuring constructs in a variance inflation factors (VIF), outer loadings, as well as the constructs’
five-point Likert scale, where 1 represented “strongly disagree” and 5 reliabilities, convergent validities, in terms of the average variance
referred to “strongly agree”. The survey was pilot tested among a small extracted (AVE) as well as their discriminant validity values.
6
Table 2
The list of measures and the corresponding items used in this research.
Construct Source Definition Code Item
Performance expectancy Unified Theory of Acceptance and Use of The performance expectancy construct is defined as PE1 ChatGPT offers a useful
Technology (UTAUT1 and UTAUT2) (Venkatesh the degree to which individuals believe that using a service.
et al., 2003; Venkatesh et al., 2012) system will help them improve their job performance. PE2 ChatGPT is convenient.
PE3 ChatGPT provides quick
answers to my questions.
PE4 ChatGPT is enhancing my
job performance.
Effort Expectancy Unified Theory of Acceptance and Use of The effort expectancy construct is defined as the EE1 It is easy to use ChatGPT.
Technology (UTAUT1 and UTAUT2) (Venkatesh degree of ease associated with the use of a system. EE2 It is not difficult to access
et al., 2003; Venkatesh et al., 2012) ChatGPT through the digital
media.
EE3 ChatGPT interacts with me
in a clear and
understandable manner.
EE4 ChatGPT is user-friendly.
Social Influences Unified Theory of Acceptance and Use of The social influences construct is defined as a process SI1 People who are important
Technology (UTAUT1 and UTAUT2) (Venkatesh that may lead individuals to change their opinions, to me think that I should use
et al., 2003; Venkatesh et al., 2012) beliefs, or behaviors as a result of social interactions ChatGPT.
with other people. SI2 People who influence my
behaviors recommend that I
use ChatGPT.
Perceived Interactivity Perceived Interactivity (McMillan and Hwang, The perceived interactivity construct is defined as the PI1 ChatGPT provides correct
2002; Zhao and Lu, 2012) individuals’ perceptions about web-based human-to- answers to my questions.
human, human-to-computer and/or computer-to- PI2 ChatGPT responds to my
human interactions, in real time. questions in real time.
Information quality Elaboration Likelihood Model (Central Route) ( The information quality construct is defined as the IQ1 The information I receive
Cacioppo and Petty, 1981), Information Adoption individuals’ perceptions about the accuracy and from ChatGPT is correct.
Model (Sussman and Siegal, 2003) reliability of the content they receive. IQ2 The information that is
provided from ChatGPT is
reliable.
Source trustworthiness Elaboration Likelihood Model (Peripheral Route) ( The source trustworthiness construct is defined as the ST1 I trust the content that is
Cacioppo and Petty, 1981); Information Adoption individuals’ perceptions about the sources’ credibility given by ChatGPT.
Model (Cheung et al., 2008) and dependability. ST2 The information I receive
from ChatGPT is
dependable.
Intentions to use the Technology Acceptance Models (TAM, TAM2 and The intentions to use the information technology INT1 I am a regular user of
information technology TAM3) (Davis et al., 1989; Davis, 1989; Venkatesh construct is defined as the individuals’ willingness to ChatGPT.
(e.g. ChatGPT) and Davis, 2000; Venkatesh and Bala, 2008); repeatedly perform specified behaviors including INT2 Most probably, I shall
Unified Theory of Acceptance and Use of utilizing information technologies (like ChatGPT). continue using ChatGPT, in
Technology (UTAUT1 and UTAUT2) (Venkatesh the near future.
et al., 2003; Venkatesh et al., 2012)
expectancy, where f2 = 0.3. Other noteworthy effects were reported

Table 3 between perceived interactivity and intentions to use ChatGPT (f2 =
The demographic profile of the research participants.
0.245), and between effort expectancy and performance expectancy (f2
Gender Age Qualifications = 0.145). There were lower effects between social influences and in
Males 292 18–28 318 Cert 14 tentions to use ChatGPT (f2 = 0.090), between performance expectancy
Females 338 29–39 110 Dip 142 and intentions to use ChatGPT (f2 = 0.057), between information quality
Other 18 40–50 128 B 194 and performance expectancy (f2 = 0.05), and between effort expectancy
51–61 74 M 162
Over 62 20 PhD 136
and intentions to use ChatGPT (f2 = 0.029). Fig. 2 depicts the path co
Preferred not to say 6 4 6 efficients of this empirical investigation.
Total (N) 654 Total (N) 654 Total (N) 654
4.3. Results from the Bootstrapping procedure

The VIFs were <3.3. The outer loadings ranged between 0.653 and
0.941. The findings confirmed that the reliability values were higher The bootstrapping procedure was utilized to examine the hypotheses
than 0.7. The AVE figures were above 0.6. The constructs’ discriminant of this study. The findings confirmed the robustness of the proposed
validities were tested through Fornell and Larcker’s (1981) criterion as structured model. They reported highly significant effects between the
well as via the HTMT procedure (Henseler et al., 2015). The former exogenous and endogenous constructs, as indicated in Table 5. The most
reported that the square roots of AVE (in bold) were higher than the significant link was found in H5, between source trustworthiness and
other correlation values (within the same columns). In addition, the performance expectancy, where β = 0.450, t = 8.477 and P < 0.001.
latter (HTMT) values, on the right-hand side of the bold figures, were Highly significant effects were reported in H7, between perceived
lower than 0.9. interactivity and intentions to use ChatGPT (β = 0.355, t = 8.255, P <
The PLS algorithm also provided details about the robustness of the 0.001), in H3, between effort expectancy and performance expectancy
structured model. It clearly indicated the factors’ predictive power and (β = 0.311, t = 6.364, P < 0.001), and in H6, between social influences
shed light on the values of R2 and ƒ2. It revealed that the independent and intentions to use ChatGPT (β = 0.263, t = 4.362, P < 0.001). Other
constructs affected 53.8 % of the users’ performance expectancy and significant effects were found in H1, between performance expectancy
61.1 % of their intentions to use ChatGPT. and intentions to use ChatGPT (β = 0.236, t = 3.029, P = 0.002), in H4,
Source trustworthiness had the highest effect on performance between information quality and performance expectancy (β = 0.158, t
= 2.966, P = 0.003), and in H2, between effort expectancy and
7
intentions to use ChatGPT (β = 0.134, t = 2.767, P = 0.006).
0.906
Table 6 summarizes the results of the mediated analyses. The find
0.587
0.528
0.631
0.749
0.839
0.43
ings reveal that performance expectancy significantly mediates effort
7 expectancy - intentions to use ChatGPT causal path. Table 7 sheds light
on the indirect effects within this research model. The results suggest
0.722
0.428
0.792
0.784
0.554
0.697
0.84
that performance expectancy significantly mediates source trustwor
6
thiness – intentions to use ChatGPT link.
0.938
0.442
0.328
0.718
0.478
0.361
0.89
5. Discussion and conclusions
5
5.1. Theoretical implications

0.931
0.634
0.422
0.873
0.614
0.669
0.62
This research sought to explore the factors that are affecting the in
4
dividuals’ intentions to use ChatGPT. It examined the online users’ effort

and performance expectancy, social influences as well as their percep
0.914
0.657
0.333
0.721
0.742
0.661
0.512
tions about the information quality, source trustworthiness and inter
3
activity of generative text AI chatbots. The empirical investigation

hypothesized that performance expectancy, effort expectancy and social
0.915
0.299
0.269
0.275
0.355
0.427
influences from Venkatesh et al.’s (2003) UTAUT together with a
0.35
2
perceived interactivity construct (McMillan and Hwang, 2002) were

significant antecedents of their intentions to revisit ChatGPT’s website
and/or to use its app. Moreover, it presumed that information quality
0.779
0.244
0.544
0.535
0.381
0.594
0.477
and source trustworthiness measures from Sussman and Siegal’s (2003)

1
IAM were found to be the precursors of performance expectancy.

The results from this study report that source trustworthiness-
0.607
0.837
0.836
0.866
0.879
0.706
0.82
AVE
performance expectancy is the most significant path in this research

model. They confirm that online users indicated that they believed that
there is a connection between the source’s trustworthiness in terms of its
0.911
0.911
0.928
0.936
0.905
0.901
dependability, and the degree to which they believe that using such an
0.86
CR
AI generative system will help them improve their job performance.

Similar effects were also evidenced in previous IAM theoretical frame
works (Kang and Namkung, 2019; Onofrei et al., 2022), as well as in a
Rho_A
0.805
0.808
0.808
0.846
0.864
0.861
0.824
number of studies related to TAM (Assaker, 2020; Chen and Aklikokou,

2020; Shahzad et al., 2018) and/or to UTAUT/UTAUT2 (Lallmahomed
et al., 2017).
Alpha
0.784
0.806
0.804
0.845
0.863
0.784
0.86
In addition, this research also reports that information quality

significantly affects their performance expectancy/expectancies from
ChatGPT. Yet, in this case, this link was weaker than the former, thus
1.802
1.429
2.045
1.835
1.835
1.825
1.825
2.156
2.156
2.358
2.358
2.817
1.886
1.636
2.752
1.713
1.713
implying that the respondents’ perceptions about the usefulness of this

1.65
VIF
text generative technology were clearly influenced by the peripheral

cues of communication (Cacioppo and Petty, 1981; Shi et al., 2018;
Sussman and Siegal, 2003; Tien et al., 2019).
Loadings
The descriptive statistics as well as the construct reliability and validity values.
0.795
0.653
0.795
0.859
0.909
0.921
0.907
0.922
0.933
0.928
0.935
0.941
0.874
0.822
0.778
0.881
0.879
0.931
Very often, academic colleagues noted that individuals would

probably rely on the information that is presented to them, if they
perceive that the sources and/or their content are trustworthy (Bingham
et al., 2019; John and De’Villiers, 2020; Winter, 2020). Frequently, they
Deviation
indicated that source trustworthiness would likely affect their beliefs

0.781
0.791
0.922
0.816
0.986
1.014
0.948
0.913
0.992
0.962
0.912
0.924
0.935
1.069
1.216
0.938
0.89
0.85
about the usefulness of information technologies, as they enable them to

enhance their performance. Conversely, some commentators argued
that there may be users that could be skeptical and wary about using
4.239
3.758
3.988
2.988
2.911
3.572
3.471
3.563
3.168
2.908
4.138
3.942
3.468
4.119
3.064
4.012
Mean
4.18
3.56
new technologies, especially if they are unfamiliar with them (Shankar

et al., 2021). They noted that such individuals may be concerned about
the reliability and trustworthiness of the latest technologies.
Items
The findings suggest that the individuals’ perceptions about the

INT1
INT2
EE1
EE2
EE3
EE4
PE1
PE2
PE3
PE4
IQ1
IQ2
IT1
IT2
PI1
PI2
SI1
SI2
interactivity of ChatGPT are a precursor of their intentions to use it. This

link is also highly significant. Therefore, the online users were somehow
appreciating this information technology’s responsiveness to their
Intentions to use ChatGPT
Performance Expectancy
prompts (in terms of its computer-human communications). Evidently,

Source trustworthiness
Perceived Interactivity
Information quality
ChatGPT’s interactivity attributes are having an impact on the in

Effort Expectancy
Social Influences
dividuals’ readiness to engage with it, and to seek answers to their

questions. Similar results were reported in other studies that analyzed
how the interactivity and anthropomorphic features of dialogue systems
like live support chatbots, or virtual assistants can influence the online
Construct
users’ willingness to continue utilizing them in the future (Baabdullah

Table 4
et al., 2022; Balakrishnan et al., 2022; Brachten et al., 2021; Liew et al.,
1
2017).
8
Fig. 2. An illustration of the results from SEM-PLS algorithm.
many individuals are favorably disposed to use dialogue systems that are
Table 5 capable of providing them with instant feedback and personalized
The findings from the Bootstrapping procedure.
content. Several colleagues suggest that positive user experiences as well
Causal path Original Standard T P as high satisfaction levels and enjoyment, could enhance their connec
sample (O) deviation statistics values
tion with information technologies, and will probably motivate them to
(STDEV)
continue using them in the future (Ashfaq et al., 2020; Camilleri and
1 Performance 0.236** 0.078 3.029 0.002 Falzon, 2021; Huang and Chueh, 2021; Wolfinbarger and Gilly, 2003).
Expectancy →
Intentions to Use
Another important finding from this research is that the individuals’
ChatGPT social influences (from family, friends or colleagues) are affecting their
2 Effort Expectancy → 0.134** 0.049 2.767 0.006 interactions with ChatGPT. Again, this causal path is also very signifi
Intentions to Use cant. Similar results were also reported in UTAUT/UTAUT2 studies that
ChatGPT
are focused on the link between social influences and its link with
3 Effort Expectancy → 0.311*** 0.047 6.634 0.000
Performance intentional behaviors to use technologies (Gursoy et al., 2019; Patil
Expectancy et al., 2020). In addition, TPB/TRA researchers found that subjective
4 Information quality → 0.158** 0.053 2.966 0.003 norms also predict behavioral intentions (Driediger and Bhatiasevi,
Performance 2019; Sohn and Kwon, 2020). This is in stark contract with other studies
Expectancy
5 Source trustworthiness 0.450*** 0.053 8.477 0.000
that reported that there was no significant relationship between social
→ Performance influences/subjective norms and behavioral intentions (Ho et al., 2020;
Expectancy Kamble et al., 2019).
6 Social Influences → 0.263*** 0.060 4.362 0.000 Interestingly, the results report that there are highly significant ef
Intentions to Use
fects between effort expectancy-performance expectancy. Many scholars
ChatGPT
7 Perceived Interactivity 0.355*** 0.043 8.255 0.000 posit that perceived ease of use is a significant driver of perceived use
→ Intentions to Use fulness of technology (Bressolles et al., 2014; Davis, 1989; Davis et al.,
ChatGPT 1989; Kamble et al., 2019; Yoo and Donthu, 2001). Furthermore, there
Note: T > 1.95. are significant causal paths between performance expectancy-intentions
***
P < 0.001. to use ChatGPT and even between effort expectancy-intentions to use
**
P < 0.01. ChatGPT, albeit to a lesser extent. Yet, this research indicates that per
formance expectancy partially mediates effort expectancy-intentions to
There are a number of academic contributions that sought to explore use ChatGPT. In this case, this link is highly significant.
how, why, where and when individuals are lured by interactive In sum, this contribution validates key information technology
communication technologies (e.g. Hari et al., 2022; Li et al., 2021; Lou measures, specifically, performance expectancy, effort expectancy, so
et al., 2022). Generally, these researchers posited that users are habit cial influences and behavioral intentions from UTAUT/UTAUT2, as well
uated with information systems that are programed to engage with them as information quality and source trustworthiness from ELM/IAM and
in a dynamic and responsive manner. Very often they indicated that integrates them with a perceived interactivity factor. It builds on
9
Table 6
The mediated effects of performance expectancy on effort expectancy-intentions to use ChatGPT.
Causal path Original Original Original Standard deviation T P Decision
sample (O) sample (O) sample (O) (STDEV) statistics values
Direct effect Indirect effect Total effects
H2 Effort Expectancy → Intentions to Use 0.134** 0.049 2.767 0.006 Supported

ChatGPT
H2a Effort Expectancy → Performance 0.073** 0.026 2.864 0.004 Supported
Expectancy → Intentions to Use ChatGPT
Effort Expectancy → Intentions to Use 0.208*** 0.047 4.454 0.000 Supported (partial
ChatGPT mediation)
Note: T > 1.95.

***
P < 0.001.
**
P < 0.01.
knowledge of the world and events after 2021 and may also occasionally
Table 7
produce harmful instructions or biased content. OpenAI recommends
The indirect effects.
checking whether its chatbot’s responses are accurate or not, and to let
Causal path Original Standard T P them know when and if it answers in an incorrect manner, by using their
sample (O) deviation statistics values
“Thumbs Down” button. They even declare that their ChatGPT’s Help
(STDEV)
Center can occasionally make up facts or “hallucinate” outputs (OpenAI,
Information quality → 0.037 0.02 1.867 0.062
2023a,b).
Performance Expectancy
→ Intentions to Use
OpenAI reports that its top notch ChatGPT Plus subscribers can ac
ChatGPT cess safer and more useful responses. In this case, users can avail
Source trustworthiness → 0.106** 0.039 2.691 0.007 themselves from a number of beta plugins and resources that can offer a
Performance Expectancy wide range of capabilities including text-to-speech applications as well
→ Intentions to Use
as web browsing features through Bing. Yet again, OpenAI (2023b) in
ChatGPT
dicates that its GPT-4 still has many known limitations that the company
Note: T > 1.95. is working to address, such as “social biases and adversarial prompts” (at
**
P < 0.01. the time of writing this article). Evidently, works are still in progress at
OpenAI. The company needs to resolve these serious issues, considering
previous theoretical underpinnings. Yet, it differentiates itself from that its Content Policy and Terms clearly stipulate that OpenAI’s con
previous studies. To date, there are no other empirical investigations sumers are the owners of the output that is created by ChatGPT. Hence,
that have combined the same constructs that are presented in this ChatGPT’s users have the right to reprint, sell, and merchandise the
article. Notwithstanding, this research puts forward a robust Informa content that is generated for them through OpenAI’s platforms,
tion Technology Acceptance Framework. The results confirm the reli regardless of whether the output (its response) was provided via a free or
ability and validity of the measures. They clearly outline the relative a paid plan.
strength and significance of the causal paths that are predicting the in Various commentators are increasingly raising awareness about the
dividuals’ intentions to use ChatGPT. corporate digital responsibilities of those involved in the research,
development and maintenance of such dialogue systems. A number of
5.2. Managerial implications stakeholders, particularly the regulatory ones, are concerned on possible
risks and perils arising from AI algorithms including interactive chat
This empirical study provides a snapshot on the online users’ per bots. In many cases, they are warning that disruptive chatbots could
ceptions about ChatGPT’s responses to verbal queries, and sheds light on disseminate misinformation, foster prejudice, bias and discrimination,
their dispositions to avail themselves from ChatGPT’s natural language raise privacy concerns, and could lead to the loss of jobs. Arguably, one
processing. It explores their performance expectations about their use has to bear in mind that, in many cases, many governments are outpaced
fulness and their effort expectations related to the ease of use of these by the proliferation of technological innovations (as their development
information technologies and investigates whether they are affected by happens before the enactment of legislation). As a result, they tend to be
colleagues or by other social influences to use such dialogue systems. reactive in the implementation of substantive regulatory interventions.
Moreover, it examines their insights about the content quality, source This research reported that the development of ChatGPT has resulted in
trustworthiness as well as on the interactivity features of these text- mixed reactions among different stakeholders in society, especially
generative AI models. during the first months after its official launch. At the moment, there are
Generally, the results suggest that the research participants felt that just a few jurisdictions that have formalized policies and governance
these algorithms are easy to use. The findings indicate that they consider frameworks that are meant to protect and safeguard individuals and
them to be useful too, specifically when the information they generate is entities from possible risks and dangers of AI technologies (Camilleri,
trustworthy and dependable. The respondents suggest that they are 2023). Of course, voluntary principles and guidelines are a step in the
concerned about the quality and accuracy of the content that is featured right direction. However, policy makers are expected by various stake
in the AI chatbots’ answers. This contingent issue can have a negative holders to step-up their commitment by introducing quasi-regulations
effect on the use of the information that is created by online dialogue and legislation.
systems. Currently, a number of technology conglomerates including
OpenAI’s ChatGPT is a case in point. Its app is freely available in Microsoft-backed OpenAI, Apple and IBM, among others, anticipated
many countries, via desktop and mobile technologies including iOS and the governments’ regulations by joining forces in a non-profit organi
Android. The company admits that its GPT-3.5 outputs may be inaccu zation entitled, “Partnership for AI” that aims to advance safe, respon
rate, untruthful, and misleading at times. It clarifies that its algorithm is sible AI, that is rooted in open innovation. In addition, IBM has also
not connected to the internet, and that it can occasionally produce teamed up with Meta and other companies, startups, universities,
incorrect answers (OpenAI, 2023a). It posits that GPT-3.5 has limited research and government organizations, as well as non-profit
10
foundations to form an “AI Alliance”, that is intended to foster in approval was obtained from the authors’ host institution.
novations across all aspects of AI technology, applications and The dataset of this research as well as all the relevant ethical docu
governance. ments are available on request.
6. Limitations and future research CRediT authorship contribution statement
This research validates measures from mainstream information Mark Anthony Camilleri: Writing – review & editing, Writing –
technology adoption models that were tried and tested in previous ac original draft, Visualization, Validation, Supervision, Software, Re
ademic literature. It utilizes performance expectancy, effort expectancy, sources, Project administration, Methodology, Investigation, Funding
social influences, behavioral intention from UTAUT/UTAUT2, infor acquisition, Formal analysis, Data curation, Conceptualization.
mation quality and source trustworthiness from ELM/IAM, as well as a
perceived interactivity construct. These seven constructs were never Declaration of competing interest
presented in the same structured model. The findings report the reli
ability and validity of the constructs used in this empirical investigation. The author declares that he has no conflicts of interest.
They indicate the robustness of the proposed theoretical framework, as
all hypotheses are supported. Hence, future researchers are invited to Data availability
replicate this study in different settings.
In the future, other scholars could rely on measures that were used in Data will be made available on request.
this study. Alternatively, they can choose other measures drawn from
extended TAM, TAM2, TAM3, TRA or TPD models, among others Acknowledgements
(featured in Table 1), to examine the individuals’ motivations to engage
with AI generative text technologies. Conversely, they may adopt spe This research was partially funded through the University of Malta’s
cific IAM constructs that examine perceptions about information quality Research Innovation and Development Trust (RSF2024-CRC-R2).
including information completeness, information accuracy, information
timeliness, information reliability, et cetera. Perhaps, prospective re
References
searchers may consider exploring other peripheral cues, including
source credibility dimensions. They could investigate the moderating Abbad, M.M., 2021. Using the UTAUT model to understand students’ usage of e-learning
effects of demographic variables, including age, gender, level of edu systems in developing countries. Educ. Inf. Technol. 26 (6), 7205–7224.
cation and occupation, among others, to better understand the in Ajzen, I., 1991. The theory of planned behavior. Organ. Behav. Hum. Decis. Process. 50
(2), 179–211.
dividuals’ dispositions to engage with AI chatbots or with other Alalwan, A.A., 2020. Mobile food ordering apps: an empirical study of the factors
interactive technologies like the Metaverse, among others. affecting customer e-satisfaction and continued intention to reuse. Int. J. Inf. Manag.
This study’s primary data was collected through a cross-sectional 50, 28–44.
Alam, M.Z., Hoque, M.R., Hu, W., Barua, Z., 2020. Factors influencing the adoption of
survey. Unlike longitudinal studies, such a research instrument pro mHealth services in a developing country: a patient-centric study. Int. J. Inf. Manag.
vides a snapshot of the research participants’ perceptions at a specific 50, 128–143.
point in time. As a result, this quantitative methodology may lend itself Al-Saedi, K., Al-Emran, M., Ramayah, T., Abusham, E., 2020. Developing a general
extended UTAUT model for M-payment adoption. Technol. Soc. 62 https://doi.org/
to possible limitations. Some colleagues argue that cross-sectional sur 10.1016/j.techsoc.2020.101293.
veys are prone to common method variance (CMV) (see Podsakoff et al., Ashfaq, M., Yun, J., Yu, S., Loureiro, S.M.C., 2020. I, Chatbot: modeling the determinants
2023). In this case, the findings confirmed that the variance inflation of users’ satisfaction and continuance intention of AI-powered service agents.
Telematics Inform. 54 https://doi.org/10.1016/j.tele.2020.101473.
factors were lower than 3.3, as per the recommended threshold (Hwang Assaker, G., 2020. Age and gender differences in online travel reviews and user-
et al., 2023). Moreover, the results reported appropriate reliability, as generated-content (UGC) adoption: extending the technology acceptance model
well as convergent and discriminant validity values. (TAM) with credibility theory. J. Hosp. Mark. Manag. 29 (4), 428–449.
Baabdullah, A.M., Alalwan, A.A., Algharabat, R.S., Metri, B., Rana, N.P., 2022. Virtual
Academic colleagues are invited to utilize other research methods
agents and flow experience: an empirical examination of AI-powered chatbots.
and sampling approaches to capture, analyze and interpret their find Technol. Forecast. Soc. Chang. 181 https://doi.org/10.1016/j.
ings. They may use inductive research designs, to reveal the research techfore.2022.12177221772.
participants’ in-depth opinions, and to evaluate their experiences with Balakrishnan, J., Abed, S.S., Jones, P., 2022. The role of meta-UTAUT factors, perceived
anthropomorphism, perceived intelligence, and social self-efficacy in chatbot-based
AI text generative technologies. Undoubtedly, this contribution is services? Technol. Forecast. Soc. Chang. 180 https://doi.org/10.1016/j.
focused on a contemporary topic in theory and practice. It is still techfore.2022.121692.
evolving and progressing, as more stakeholders are devoting their en Bauer, H.H., Falk, T., Hammerschmidt, M., 2006. eTransQual: a transaction process-
based approach for capturing service quality in online shopping. J. Bus. Res. 59 (7),
ergies to continue improving the quality of LLMs. In this light, there is 866–875.
scope for researchers to continue investigating conversational (verbal) Beh, P.K., Ganesan, Y., Iranmanesh, M., Foroughi, B., 2021. Using smartwatches for
capabilities as well as the anthropomorphic (visual and vocal) features fitness and health monitoring: the UTAUT2 combined with threat appraisal as
moderators. Behav. Inform. Technol. 40 (3), 282–299.
of chatbots. Besides, they are also urged to explore the governments’ Bingham, C.B., Howell, T., Ott, T.E., 2019. Capability creation: heuristics as
regulatory and quasi-regulatory interventions (to shed light on their microfoundations. Strateg. Entrep. J. 13 (2), 121–153.
principles, soft and hard laws) in this regard. Brachten, F., Kissmer, T., Stieglitz, S., 2021. The acceptance of chatbots in an enterprise
context–a survey study. Int. J. Inf. Manag. 60 https://doi.org/10.1016/j.
ijinfomgt.2021.102375.
Author’s statement Bressolles, G., Durrieu, F., Senecal, S., 2014. A consumer typology based on e-service
quality and e-satisfaction. J. Retail. Consum. Serv. 21 (6), 889–896.
Cacioppo, J.T., Petty, R.E., 1981. Electromyograms as measures of extent and affectivity
The corresponding (sole) author of this manuscript confirms that he
of information processing. Am. Psych. 36 (5), 441.
has no conflicts of interest and that all statements presented in this Camilleri, M.A., 2023. Artificial intelligence governance: ethical considerations and
manuscript are correct. implications for social responsibility. Expert. Syst., e13406. https://onlinelibrary.
wiley.com/doi/epdf/10.1111/exsy.1340.
Camilleri, M.A., Camilleri, A.C., 2022. The functionality and interactivity of video
Research ethics conferencing technologies: a synchronous technology adoption model. In:
Proceedings of the 2022 5th International Conference on Education Technology
This research was carried out in accordance with the principles Management. New York, USA, ACM Digital Library, pp. 199–204.
Camilleri, M.A., Falzon, L., 2021. Understanding motivations to use online streaming
stated in the Declaration of Helsinki and is congruent with the European services: integrating the technology acceptance model (TAM) and the uses and
Union’s General Data Protection Regulations (EU, 2016). Ethical gratifications theory (UGT). Span. J. Mark. - ESIC 25 (2), 217–238.
11
Camilleri, M.A., Filieri, R., 2023. Customer satisfaction and loyalty with online consumer Kang, J.W., Namkung, Y., 2019. The information quality and source credibility matter in
reviews: factors affecting revisit intentions. Int. J. Hosp. Manag. 114 https://doi.org/ customers’ evaluation toward food O2O commerce. Int. J. Hosp. Manag. 78,
10.1016/j.ijhm.2023.103575. 189–198.
Camilleri, M.A., Kozak, M., 2022. Interactive engagement through travel and tourism Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F.,
social media groups: a social facilitation theory perspective. Technol. Soc. 71 Kasneci, G., 2023. ChatGPT for good? On opportunities and challenges of large
https://doi.org/10.1016/j.techsoc.2022.102098. language models for education. Learn. Individ. Differ. 103 https://doi.org/10.1016/
Camilleri, M.A., Kozak, M., 2023. Utilitarian motivations to engage with travel websites: j.lindif.2023.102274.
an interactive technology adoption model. J. Serv. Mark. 37 (1), 96–109. Kaya, B., Behravesh, E., Abubakar, A.M., Kaya, O.S., Orús, C., 2019. The moderating role
Camilleri, M.A., Troise, C., Kozak, M., 2023. Functionality and usability features of of website familiarity in the relationships between e-service quality, e-satisfaction
ubiquitous mobile technologies: the acceptance of interactive travel apps. J. Hosp. and e-loyalty. J. Internet Commer. 18 (4), 369–394.
Tour. Technol. 14 (2), 188–207. Lallmahomed, M.Z., Lallmahomed, N., Lallmahomed, G.M., 2017. Factors influencing
Chen, L., Aklikokou, A.K., 2020. Determinants of E-government adoption: testing the the adoption of e-government services in Mauritius. Telematics Inform. 34 (4),
mediating effects of perceived usefulness and perceived ease of use. Int. J. Public 57–72.
Adm. 43 (10), 850–865. Leong, C.M., Loi, A.M.W., Woon, S., 2021. The influence of social media eWOM
Cheung, C.M., Lee, M.K., Rabjohn, N., 2008. The impact of electronic word-of-mouth: the information on purchase intention. J. Mark. Anal. 10, 145–157.
adoption of online opinions in online customer communities. Internet Res. 18 (3), Li, L., Lee, K.Y., Emokpae, E., 2021. What makes you continuously use chatbot services?
229–247. Evidence from chinese online travel agencies. Electron. Mark. 31, 575–599.
Compeau, D.R., Higgins, C.A., 1995. Application of social cognitive theory to training for Liebrenz, M., Schleifer, R., Buadze, A., Bhugra, D., Smith, A., 2023. Generating scholarly
computer skills. Inf. Syst. Res. 6 (2), 118–143. content with ChatGPT: ethical challenges for medical publishing. Lancet Digit.
Cortez, R.M., Johnston, W.J., 2020. The Coronavirus crisis in B2B settings: crisis Health 5 (3), e105–e106.
uniqueness and managerial implications based on social exchange theory. Ind. Mark. Liew, T.W., Tan, S.M., Ismail, H., 2017. Exploring the effects of a non-interactive talking
Manag. 88, 125–135. avatar on social presence, credibility, trust, and patronage intention in an e-
Davis, F.D., 1989. Perceived usefulness, perceived ease of use, and user acceptance of commerce website. HCIS 7, 1–21.
information technology. MIS Q. 13 (3), 319–340. Lou, C., Kang, H., Tse, C.H., 2022. Bots vs. humans: how schema congruity, contingency-
Davis, F.D., Bagozzi, R.P., Warshaw, P.R., 1989. User acceptance of computer based interactivity, and sympathy influence consumer perceptions and patronage
technology: a comparison of two theoretical models. Manag. Sci. 35 (8), 982–1003. intentions. Int. J. Advert. 41 (4), 655–684.
Davis, F.D., Bagozzi, R.P., Warshaw, P.R., 1992. Extrinsic and intrinsic motivation to use Luqman, A., Zhang, Q., Hina, M., 2023. Employees’ proactiveness on enterprise social
computers in the workplace. J. Appl. Soc. Psychol. 22, 1111–1132. media and social consequences: an integrated perspective of social network and
Driediger, F., Bhatiasevi, V., 2019. Online grocery shopping in Thailand: consumer social exchange theories. Inf. Manag. https://doi.org/10.1016/j.im.2023.103843.
acceptance and usage behavior. J. Retail. Consum. Serv. 48, 224–237. Malodia, S., Islam, N., Kaur, P., Dhir, A., 2021. Why do people use artificial intelligence
Erkan, I., Evans, C., 2016. The influence of eWOM in social media on consumers’ (AI)-enabled voice assistants? IEEE Trans. Eng. Manag. https://doi.org/10.1109/
purchase intentions: an extended approach to information adoption. Comp. Human TEM.2021.3117884.
Behav. 61, 47–55. McMillan, S.J., Hwang, J.S., 2002. Measures of perceived interactivity: an exploration of
Erkan, I., Evans, C., 2018. Social media or shopping websites? The influence of eWOM on the role of direction of communication, user control, and time in shaping perceptions
consumers’ online purchase intentions. J. Mark. Commun. 24 (6), 617–632. of interactivity. J. Advert. 31 (3), 29–42.
EU, 2016. Regulation (EU) 2016/679 of the European Parliament and of the Council of MIT, 2023. ChatGPT is going to change education, not destroy it. MIT Technol. Rev.
27 April 2016 on the Protection of Natural Persons With Regard to the Processing of https://www.technologyreview.com/2023/04/06/1071059/chatgpt-change-not-de
Personal Data and on the Free Movement of Such Data, and Repealing Directive 95/ stroy-education-openai/.
46/EC (General Data Protection Regulation) (Text With EEA Relevance). European Molinillo, S., Anaya-Sánchez, R., Liebana-Cabanillas, F., 2020. Analyzing the effect of
Commission, Brussels, Belgium. https://eur-lex.europa.eu/eli/reg/2016/679/oj. social support and community factors on customer engagement and its impact on
Farrokhnia, M., Banihashem, S.K., Noroozi, O., Wals, A., 2023. A SWOT analysis of loyalty behaviors toward social commerce websites. Comput. Hum. Behav. 108
ChatGPT: implications for educational practice and research. Innov. Educ. Teach. https://doi.org/10.1016/j.chb.2019.04.004.
Int. https://doi.org/10.1080/14703297.2023.2195846. Moore, G.C., Benbasat, I., 1991. Development of an instrument to measure the
Ferguson, J.L., Mohan, M., 2020. Use of celebrity and non-celebrity persons in B2B perceptions of adopting an information technology innovation. Inf. Syst. Res. 2 (3),
advertisements: effects on attention, recall, and hedonic and utilitarian attitudes. 192–222.
Ind. Mark. Manag. 89, 594–604. OECD, 2023. AI Language Models: Technological, Socio-economic and Policy
Fishbein, M., Ajzen, I., 1975. Predicting and understanding consumer behavior: attitude- Considerations. Organization for Economic Cooperation and Development, Paris,
behavior correspondence. In: Ajzen, I., Fishbein, M. (Eds.), Understanding Attitudes France. https://www.oecd.org/publications/ai-language-models-13d38f92-en.htm.
and Predicting Social Behavior. Prentice Hall, Englewood Cliffs, NJ, pp. 148–172. Onofrei, G., Filieri, R., Kennedy, L., 2022. Social media interactions, purchase intention,
Fornell, C., Larcker, D.F., 1981. Evaluating structural equation models with unobservable and behavioural engagement: the mediating role of source and content factors.
variables and measurement error. J. Mark. Res. 18 (1), 39–50. J. Bus. Res. 142, 100–112.
Gill, S.S., Xu, M., Patros, P., Wu, H., Kaur, R., Kaur, K., Buyya, R., 2024. Transformative OpenAI, 2023a. What Is ChatGPT. Commonly Asked Questions About ChatGPT. OpenAI,
effects of ChatGPT on modern education: emerging era of AI Chatbots. Internet of San Francisco, California, United States of America. https://help.openai.com/en/arti
Things and Cyber-Phys. Syst. 4, 19–23. cles/6783457-what-is-chatgpt.
Gursoy, D., Chi, O.H., Lu, L., Nunkoo, R., 2019. Consumers acceptance of artificially OpenAI, 2023b. OpenAI’s Most Advanced System, Producing Safer Andmore Useful
intelligent (AI) device use in service delivery. Int. J. Inf. Manag. 49, 157–169. Responses. OpenAI, San Francisco, California, United States of America. https
Hari, H., Iyer, R., Sampat, B., 2022. Customer brand engagement through chatbots on ://openai.com/gpt-4.
bank websites–examining the antecedents and consequences. Int. J. Hum.-Comput. Parasuraman, A., Zeithaml, V.A., Malhotra, A., 2005. ES-QUAL: a multiple-item scale for
Interact. 38 (13), 1212–1227. assessing electronic service quality. J. Serv. Res. 7 (3), 213–233.
Henseler, J., Ringle, C.M., Sarstedt, M., 2015. A new criterion for assessing discriminant Patil, P., Tamilmani, K., Rana, N.P., Raghavan, V., 2020. Understanding consumer
validity in variance-based structural equation modeling. J. Acad. Mark. Sci. 43, adoption of mobile payment in India: extending Meta-UTAUT model with personal
115–135. innovativeness, anxiety, trust, and grievance redressal. Int. J. Inf. Manag. 54 https://
Ho, J.C., Wu, C.G., Lee, C.S., Pham, T.T.T., 2020. Factors affecting the behavioral doi.org/10.1016/j.ijinfomgt.2020.102144.
intention to adopt mobile banking: an international comparison. Technol. Soc. 63, Peltier, J.W., Dahl, A.J., Schibrowsky, J.A., 2023. Artificial intelligence in interactive
101360. marketing: a conceptual framework and research agenda. J. Res. Interact. Mark.
Hoeken, H., Hornikx, J., Linders, Y., 2020. The importance and use of normative criteria https://doi.org/10.1108/JRIM-01-2023-0030.
to manipulate argument quality. J. Advert. 49 (2), 195–201. Peng, L., Liao, Q., Wang, X., He, X., 2016. Factors affecting female user information
Huang, D.H., Chueh, H.E., 2021. Chatbot usage intention analysis: veterinary adoption: an empirical investigation on fashion shopping guide websites. Electron.
consultation. J. Innov. Knowl. 6 (3), 135–144. Commer. Res. 16, 145–169.
Hwang, H., Sarstedt, M., Cho, G., Choo, H., Ringle, C.M., 2023. A primer on integrated Podsakoff, P.M., Podsakoff, N.P., Williams, L.J., Huang, C., Yang, J., 2023. Common
generalized structured component analysis. Eur. Bus. Rev. 35 (3), 261–284. method bias: it’s bad, it’s complex, it’s widespread, and it’s not easy to fix. Annu.
Jiang, Y., Yang, X., Zheng, T., 2023. Make chatbots more adaptive: dual pathways linking Rev. Organ. Psych. Organ. Behav. 11. https://www.annualreviews.org/doi/abs/
human-like cues and tailored response to trust in interactions with chatbots. 10.1146/annurev-orgpsych-110721-040030.
Comput. Hum. Behav. 138 https://doi.org/10.1016/j.chb.2022.107485. Queiroz, M.M., Fosso Wamba, S., De Bourmont, M., Telles, R., 2021. Blockchain adoption
Jin, X.L., Cheung, C.M., Lee, M.K., Chen, H.P., 2009. How to keep members using the in operations and supply chain management: empirical evidence from an emerging
information in a computer-supported social network. Comput. Hum. Behav. 25 (5), economy. Int. J. Prod. Res. 59 (20), 6087–6103.
1172–1181. Qureshi, R., Shaughnessy, D., Gill, K.A.R., 2023. Are ChatGPT and large language models
John, S.P., De’Villiers, R., 2020. Elaboration of marketing communication through visual “the answer” to bringing us closer to systematic review automation? Syst. Rev. 12,
media: an empirical analysis. J. Retail. Consum. Serv. 54 https://doi.org/10.1016/j. 72 https://doi.org/10.1186/s13643-023-02243-z.
jretconser.2020.102052. Raza, S.A., Qazi, W., Khan, K.A., Salam, J., 2021. Social isolation and acceptance of the
Kamal, S.A., Shafiq, M., Kakria, P., 2020. Investigating acceptance of telemedicine learning management system (LMS) in the time of COVID-19 pandemic: an
services through an extended technology acceptance model (TAM). Technol. Soc. 60 expansion of the UTAUT model. J. Educ. Comput. Res. 59 (2), 183–208.
https://doi.org/10.1016/j.techsoc.2019.101212. Reuters, 2023. Italy restores ChatGPT after OpenAI responds to regulator. https://www.
Kamble, S., Gunasekaran, A., Arha, H., 2019. Understanding the Blockchain technology reuters.com/technology/chatgpt-is-available-again-users-italy-spokesperson-says-20
adoption in supply chains-Indian context. Int. J. Prod. Res. 57 (7), 2009–2033. 23-04-28/.
12
Rohde, P., Mau, G., 2021. “It’s selling like hotcakes”: deconstructing social media Venkatesh, V., Morris, M.G., Davis, G.B., Davis, F.D., 2003. User acceptance of
influencer marketing in long-form video content on youtube via social influence information technology: toward a unified view. MIS Q. 27 (3), 425–478.
heuristics. Eur. J. Mark. 55 (10), 2700–2734. Venkatesh, V., Thong, J.Y., Xu, X., 2012. Consumer acceptance and use of information
Rudolph, J., Tan, S., Tan, S., 2023. ChatGPT: bullshit spewer or the end of traditional technology: extending the unified theory of acceptance and use of technology. MIS
assessments in higher education? J. Appl. Learn. Teach. 6 (1), 342–363. Q. 36 (1), 157–178.
Shahzad, F., Xiu, G., Wang, J., Shahbaz, M., 2018. An empirical investigation on the Winter, S., 2020. Do anticipated Facebook discussions diminish the importance of
adoption of cryptocurrencies among the people of mainland China. Technol. Soc. 55, argument quality? An experimental investigation of attitude formation in social
33–40. media. Media Psychol. 23 (1), 79–106.
Shankar, V., Kalyanam, K., Setia, P., Golmohammadi, A., Tirunillai, S., Douglass, T., Wolfinbarger, M., Gilly, M.C., 2003. eTailQ: dimensionalizing, measuring and predicting
Waddoups, R., 2021. How technology is changing retail. J. Retail. 97 (1), 13–27. etail quality. J. Retail. 79 (3), 183–198.
Sharma, S., Islam, N., Singh, G., Dhir, A., 2022. Why do retail customers adopt artificial Yoo, B., Donthu, N., 2001. Developing a scale to measure the perceived quality of an
intelligence (Ai) based autonomous decision-making systems? IEEE Trans. Eng. Internet shopping site (SITEQUAL). Q. J. Electron. Commer. 2 (1), 31–45.
Manag. https://doi.org/10.1109/TEM.2022.3157976. Zhang, H., Lu, Y., Gupta, S., Zhao, L., 2014. What motivates customers to participate in
Shi, J., Hu, P., Lai, K.K., Chen, G., 2018. Determinants of users’ information social commerce? The impact of technological environments and virtual customer
dissemination behavior on social networking sites: an elaboration likelihood model experiences. Inf. Manag. 51 (8), 1017–1030.
perspective. Internet Res. 28 (2), 393–418. Zhao, L., Lu, Y., 2012. Enhancing perceived interactivity through network externalities:
Sohn, K., Kwon, O., 2020. Technology acceptance theories and factors influencing an empirical study on micro-blogging service satisfaction and continuance intention.
artificial intelligence-based intelligent products. Telematics Inform. 47 https://doi. Decis. Support. Syst. 53 (4), 825–834.
org/10.1016/j.tele.2019.101324. Zhao, Y., Bacao, F., 2020. What factors determining customer continuingly using food
Song, J.H., Zinkhan, G.M., 2008. Determinants of perceived web site interactivity. delivery apps during 2019 novel coronavirus pandemic period? Int. J. Hosp. Manag.
J. Mark. 72 (2), 99–113. 91 https://doi.org/10.1016/j.ijhm.2020.102683.
Sussman, S.W., Siegal, W.S., 2003. Informational influence in organizations: an
integrated approach to knowledge adoption. Inf. Syst. Res. 14 (1), 47–65.
Mark Anthony Camilleri, Ph.D. (Edinburgh) is an Associate Professor in the Department
Takefuji, Y., 2023. A brief tutorial on generative AI. Br. Dent. J. 234 (12), 845 https://
of Corporate Communication at the University of Malta, Malta. He was awarded a Ful
doi.org/10.1038/s41415-023-6041-0.
bright Visiting Academic Scholarship at NorthWestern University in Evanston, USA. Prof.
Tam, C., Santos, D., Oliveira, T., 2020. Exploring the influential factors of continuance
Camilleri was featured among the world’s top 2 % scientists in an author database of
intention to use mobile apps: extending the expectation confirmation model. Inf.
standardized citation indicators, published through Elsevier’s Mendeley Data (in 2021 and
Syst. Front. 22, 243–257.
2022).
Thompson, R.L., Higgins, C.A., Howell, J.M., 1991. Personal computing: toward a
He holds a Ph.D. (Management) from the University of Edinburgh in Scotland, an MBA
conceptual model of utilization. MIS Q. 15 (1), 125–143.
from the University of Leicester, England, and an M.Sc. from the University of Portsmouth,
Thorp, H.H., 2023. ChatGPT is fun, but not an author. Science 379 (6630), 313. https://
England. Prof. Camilleri is regularly engaged as a scientific expert, reviewer or as a foreign
www.science.org/doi/full/10.1126/science.adg7879.
member of expert teams of various (national) research councils in Europe. He has been
Tien, D.H., Rivas, A.A.A., Liao, Y.K., 2019. Examining the influence of customer-to-
recognized as an “excellent reviewer” as well as a “top peer reviewer” by Publons. In 2022,
customer electronic word-of-mouth on purchase intention in social networking sites.
he was awarded a Literati Award by Emerald, for his “outstanding reviews”.
Asia Pac. Manag. Rev. 24 (3), 238–249.
He is an Associate Editor of Business Strategy and the Environment and of Interna
Venkatesh, V., Davis, F.D., 2000. A theoretical extension of the technology acceptance
tional Journal of Hospitality Management, among others.
model: four longitudinal field studies. Manag. Sci. 46 (2), 186–204.
Venkatesh, V., Bala, H., 2008. Technology acceptance model 3 and a research agenda on
interventions. Decis. Sci. 39 (2), 273–315.
13

1 s2.0 S004016252400043X Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S004016252400043X Main

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

1 s2.0 S004016252400043X Main

Uploaded by

Copyright:

Available Formats

Technological Forecasting & Social Change 201 (2024) 123247

Contents lists available at ScienceDirect

Technological Forecasting & Social Change

Factors affecting performance expectancy and intentions to use ChatGPT:

Fig. 1. Information technology acceptance framework.

expectancy, where f2 = 0.3. Other noteworthy effects were reported

4.3. Results from the Bootstrapping procedure

intentions to use ChatGPT (β = 0.134, t = 2.767, P = 0.006).

thiness – intentions to use ChatGPT link.

5.1. Theoretical implications

dividuals’ intentions to use ChatGPT. It examined the online users’ effort

activity of generative text AI chatbots. The empirical investigation

perceived interactivity construct (McMillan and Hwang, 2002) were

and source trustworthiness measures from Sussman and Siegal’s (2003)

IAM were found to be the precursors of performance expectancy.

performance expectancy is the most significant path in this research

AI generative system will help them improve their job performance.

number of studies related to TAM (Assaker, 2020; Chen and Aklikokou,

In addition, this research also reports that information quality

implying that the respondents’ perceptions about the usefulness of this

text generative technology were clearly influenced by the peripheral

Very often, academic colleagues noted that individuals would

indicated that source trustworthiness would likely affect their beliefs

about the usefulness of information technologies, as they enable them to

new technologies, especially if they are unfamiliar with them (Shankar

The findings suggest that the individuals’ perceptions about the

interactivity of ChatGPT are a precursor of their intentions to use it. This

prompts (in terms of its computer-human communications). Evidently,

ChatGPT’s interactivity attributes are having an impact on the in­

dividuals’ readiness to engage with it, and to seek answers to their

users’ willingness to continue utilizing them in the future (Baabdullah

Fig. 2. An illustration of the results from SEM-PLS algorithm.

Direct effect Indirect effect Total effects

H2 Effort Expectancy → Intentions to Use 0.134** 0.049 2.767 0.006 Supported

Note: T > 1.95.

6. Limitations and future research CRediT authorship contribution statement

You might also like

ChatGPT’s interactivity attributes are having an impact on the in