Content uploaded by Md. Rabiul Awal
Author content
All content in this area was uploaded by Md. Rabiul Awal on Apr 02, 2024
Content may be subject to copyright.
Revisiting university students’
intention to accept AI-Powered
chatbot with an integration
between TAM and SCT: a south
Asian perspective
Md. Rabiul Awal
Department of Business Administration,
Bangladesh Army University of Science and Technology, Saidpur, Bangladesh, and
Md. Enamul Haque
Department of Management,
Bangamata Sheikh Fojilatunnesa Mujib Science and Technology University,
Jamalpur, Bangladesh
Abstract
Purpose –This paper aims to explore students’intention to use and actual use of the artificial intelligence (AI)-
based chatbot such as ChatGPT or Google Bird in the field of higher education in an emerging economic context
like Bangladesh.
Design/methodology/approach –The present study uses convenience sampling techniques to collect data
from the respondents. It applies partial least squares structural equation modeling (PLS-SEM) for analyzing a
total of 413 responses to examine the study’s measurement and structural model.
Findings –The results explore that perceived ease of use (PEOU) negatively affects intention to adopt AI-
powered chatbots (IA), whereas university students’perceived usefulness (PU) influences their IA positively
but insignificantly. Furthermore, time-saving feature (TSF), academic self-efficacy (ASE) and electronic word-
of-mouth (EWOM) have a positive and direct impact on their IA. The finding also reveals that students’IA
positively and significantly affects their actual use of AI-based chatbot (AU). Precisely, out of the five
constructs, the TSF has the strongest impact on students’intentions to use chatbots.
Practical implications –Students who are not aware of the chatbot usage benefits might ignore these
AI-powered language models. On the other hand, developers of chatbots may not be conscious of the crucial
drawbacks of their product as per the perceptions of their multiple users. However, the findings transmit a clear
message about advantages to users and drawbacks to developers. Therefore, the results will enhance the
chatbots’functionality and usage.
Originality/value –The findings of the study alert the teachers, students and policymakers of higher
educational institutions to understand the positive outcomes and to accept AI-powered chatbots such as
OpenAI’s ChatGPT. Outcomes also notify the AI-product developers to boost the chatbot’s quality in terms of
timeliness, user-friendliness, accuracy and trustworthiness.
Keywords Artificial intelligence, Chatbot, ChatGPT, TAM, SCT, Bangladesh
Paper type Research paper
Students’
intention to
accept AI-
based chatbot
The authors provide heartfelt gratitude and thanks to all experts who have assisted in conducting pilot
testing and all respondents who have given their valuable responses timely.
Funding: This innovative project has not received financial support from either governmental or
private entities. Furthermore, the author’s affiliated institutions have not designated funding for this
quantitative endeavor.
Future research direction: The current study only focuses on investigating the university students’
behavioral intention toward the adoption of AI-powered chatbots. Furthermore, this paper only applies
the original TAM research paradigm to integrate with SCT for developing the conceptual framework.
Therefore, future researchers might emphasize more rigorous studies to explore university teachers’
behavioral responses toward adopting digital chatbots by applying extended TAM or unified theory of
acceptance and use of technology (UTAUT).
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/2050-7003.htm
Received 9 November 2023
Revised 30 December 2023
14 February 2024
Accepted 18 February 2024
Journal of Applied Research in
Higher Education
© Emerald Publishing Limited
2050-7003
DOI 10.1108/JARHE-11-2023-0514
1. Introduction
In the past 7 decades, scientists and scholars globally conducted groundbreaking researches
on artificial intelligence (AI), which involves developing AI comparable to humans, utilizing
computer and Internet technologies. AI significantly impacts various aspects of daily life,
including the economy, healthcare, business, politics and education (Jiang et al., 2022). With
the remarkable success of ChatGPT, AI has gained immense attention across platforms,
emerging as a prominent research trend (Holzinger et al., 2023).
Academia is undergoing digital transformation, aligning with the fourth industrial
revolution seen in healthcare, businesses and transportation. The focus now shifts to
debating whether AI-based digital chatbots, like ChatGPT, are a curse or blessing for
students and faculties. Developed by OpenAI, a San Francisco-based tech firm, ChatGPT
evolves from version 3 to version 4 (Biswas, 2023).
As an emerging economic country, Bangladesh witnesses growing interest among
students from various higher educational institutions in using ChatGPT since its release on
November 30, 2022. However, no significant studies, especially in education, explored the
intentions and actual use of ChatGPT among students in Bangladesh. To address these gaps,
the present study aims to set objectives through data collection and analysis, requiring
advanced strategic plans, tactics and policies for adopting new technologies like ChatGPT
(Awal et al., 2023a,b). The developed objectives are as follows:
(1) To explore the impact of socio-psychological factors [such as perceived ease of use
(PEOU), perceived usefulness (PU), time-saving feature (TSF), electronic word-of-
mouth (EWOM) and academic self-efficacy (ASE)] on university students’intention to
use AI-based chatbots (IA) in the context of Bangladesh.
(2) To reveal the impact of university students’intention to use AI-powered chatbots (IA)
on the actual use of AI-based chatbots (AU) in the context of Bangladesh.
(3) To validate the integration between the technology acceptance model (TAM) with
social cognitive theory (SCT).
The outcome of this comprehensive work will certainly assist the developers of multiple
chatbots such as ChatGPT or Google Bird to increase the functionality with the minimum
time required to respond to complex queries.
2. Background study and theoretical underpinning
2.1 Chatbot foundation
Chatbots were developed in 1950 but in recent years tech firms have grabbed this natural
language model to gain better insights into their target customers (Kaczorowska-Spychalska,
2019). As a hot trend, the present study chooses AI-powered chatbot (ChatGPT) as the study
domain.
2.2 Technology acceptance model (TAM)
The TAM, developed by Davis (1989) and modified from Fishbein and Azen’s (1975) Theory
of Reasoned Action, predicts users’technology acceptance (see Figure 1). Davis introduced
PEOU and PU as key predictors for investigating users’behavioral attitude and intention to
use new technology. PEOU reflects users expecting a hassle-free and user-friendly experience
(Na et al., 2022), while PU signifies users anticipating significant opportunities and improved
performance (Bailey et al., 2022). This study applies TAM to interpret and predict
Bangladeshi university students’intention and actual use of AI-based ChatGPT (Bin-
Nashwan et al., 2023).
JARHE
2.3 Social cognitive theory (SCT)
Rana and Dwivedi (2015) applied SCT to investigate users’intentions in adopting new
technologies like e-banking, e-government and digital platforms. SCT considered cognitive
factors (self-efficacy, values), environmental factors and stress, impacting decision-making in
accepting new technology (Bandura, 2023). This study examines TSF, ASE and EWOM as
independent variables, exploring Bangladeshi university students’behavioral intention and
use of AI-based ChatGPT (Bin-Nashwan et al., 2023).
2.4 Integration between TAM and SCT
The present study uses TAM by Davis (1989) through an integration with SCT by Bandura
(1986) to support the conceptual framework of the study. To the best of the author’s
knowledge, no single study has utilized the integration of TAM and SCT to investigate the
intention and real-world application of AI-based ChatGPT by university students in
Bangladesh and around the world. Meanwhile, this novel theoretical integration was applied
to reveal the clients’mobile banking adoption intention and actual use (Hajiyev and Chang,
2017). Theories centered on human behavior suggest that an individual’s inclination to
engage in a particular action or embrace a new aspect is influenced by psychological,
behavioral, environmental and cognitive factors related to their lifestyle. So, the literature
rationally supports and positively validates the integration between TAM and SCT to
conduct comprehensive research on exploring students’ChatGPT adoption intention and
actual use (Davis, 1989;Lee et al., 2023).
2.5 Perceived ease of use (PEOU) and intention to adopt AI-based chatbot
As per Davis (1989), perceived ease of use (PEOU) entails users’perceptions of how effortless
it is to utilize a new technology or system, aiming for minimal effort. Hamidi and Chavoshi
(2018) asserted that researchers are deeply engaged in pioneering research and experiments
to simplify people’s lives by creating innovative technology employing computers and
information systems. Meanwhile, a comprehensive study by Tiwari et al. (2023) revealed a
highly paradoxical result that PEOU has no significant positive impact on students’adoption
intention of AI-based ChatGPT. Hence, given the mixed outcomes in this field, a more
comprehensive investigation is deemed necessary. Consequently, the authors posit the
following correlation between PEOU and the inclination to adopt AI-based ChatGPT, aiming
to elucidate the unresolved paradox.
H1. PEOU does not positively influence the university students’behavioral IA.
2.6 Perceived usefulness (PU) and intention to adopt AI-based chatbot
PU as one of the constructs of TAM transparently depicts the system users’degree of trust
that a newly developed system might boost the effectiveness, performance and quality of
their job (Davis, 1989). Numerous prior studies have delved into examining the impact of
PU on users’behavioral intentions to adopt technology. According to Astuti (2023),the
Figure 1.
TAM Model developed
by Davis (1989)
Students’
intention to
accept AI-
based chatbot
extent of belief among bank clients in the utility of adopting Internet banking serves as a
reliable predictor for their inclination to use i-banking. Therefore, the authors accept that
university students’IA highly depends on their PU. Thus, this study posits the following
association between PU and students’behavioral IA as the second hypothesis of the
current study:
H2. PU significantly influences the university students’behavioral IA.
2.7 Time-saving feature and intention to adopt AI-based chatbots
In the era of Industry 4.0, time is considered the most precious resource, as individuals feel
compelled to complete multiple tasks within a 24-h period. Martha (2009) highlighted that
human psychology, particularly the perception of using time effectively, efficiently and
significantly influences behavior and related intentions in social and environmental contexts.
Ng et al. (2023) further emphasized that the adoption intention of technology users is
significantly shaped by the timely features associated with using technology. Therefore, the
current study spells out based on the above-discussed literature, university students and
faculty members might display positive intention to accept AI-based chatbots of chatbot’s
timeline feature. Thus, this study posits the following hypothesis:
H3. TSF positively influences the university students’behavioral IA.
2.8 Electronic word-of-mouth (EWOM) and intention to adopt AI-based chatbot
EWOM is the online expression of reviews and unbiased opinions on products or services by
experienced customers, aiding new buyers on platforms (Li et al., 2022). Literature on
EWOM’s direct impact on students’AI-based chatbot adoption is limited. Previous studies
focused on EWOM’s influence on trust, perception, satisfaction and buying attitude
(Kusawat and Teerakapibal, 2022). Bin-Nashwan et al. (2023) found a positive impact of
chatbot users’EWOM on their willingness to adopt AI-based chatbots like OpenAI’s
ChatGPT for academic and non-academic tasks. However, more rigorous studies are needed
for stronger theoretical evidence, addressing queries of readers, academics and
policymakers. This study posits a hypothesis on the relationship between EWOM and
the IA.
H4. EWOM positively influences the university students’behavioral IA.
2.9 Academic self-efficacy and intention to adopt AI-based chatbot
Self-efficacy, akin to self-confidence, reflects one’s belief in their ability. Academic self-
efficacy gauges a student’s confidence in task performance or learning new material
(O’Connor and Mahony, 2023). ChatGPT’s launch transformed academic approaches, offered
continuous support for information retrieval and problem-solving, enhanced self-efficacy for
both teachers and students (Rudolph et al., 2023). Additionally, Bin-Nashwan et al. (2023)
noted a positive link between academic self-efficacy and users’intention to use AI-based
ChatGPT.
Therefore, the present study develops the following relationship as the hypothesis of the
current study:
H5. Academic self-efficacy positively influences the university students’behavioral IA.
2.10 Intention to adopt AI-based chatbot and actual use of chatbot
The acceptance of new elements by users is steered by their behavioral intention and the
actual usage serves as the manifestation of those intentions. Current research underscores the
JARHE
positive influence of behavioral intention on the utilization of systems. Raman et al. (2022)
found a positive link between users’adoption intention and actual usage in a study on
learning management systems. Similarly, Awal et al. (2023a,b) discovered online customers’
future buying willingness influenced belief in fraudulent news. This study endeavors to fill
the existing gap in comprehending the connection between university students’inclination to
utilize AI-based chatbots and their real usage. The hypothesis formulated for investigation is
as follows:
H6. IA positively influences the university students’AU.
The study explores global perspectives on AI-powered chatbots, revealing challenges
identified by Lin et al. (2023) in a 23-year systematic review. Key issues include aligning user
queries with language models for precise answers and ensuring user readiness for AI-
powered chatbots (Chhibber and Bhadauria, 2022).
The literature review reveals clear gaps in understanding: mixed results on PEOU’s
impact on students’AI-powered chatbot adoption intention, a country gap in Bangladeshi
context, a theoretical gap in TAM-SCT integration and a lack of studies on the consequence of
adoption intention on actual usage. This study proposes a research model addressing these
gaps (see Figure 2).
Uniqueness is highly required for all kinds of research, whether it may be from the social
science domain or any other domain. The present study has a particular uniqueness since it
blends two renowned theoretical paradigms including TAM and SCT to investigate the
students’willingness to accept and use AI-powered chatbots for the very first time.
3. Methodology
3.1 Sampling frame and survey instrument
The study utilizes convenience sampling as like other social science researches (Alam, 2022).
This non-probability technique involves selecting easily accessible and communicative
respondents (Liu et al., 2022). The research targets university students in Bangladesh who
have used an AI-powered chatbot like ChatGPT at least once. Major universities in
Bangladesh like, Dhaka University, Rajshahi University, North South University and
Bangladesh Army University of Science and Technology are included for conducting the
survey. Survey instruments are developed based on existing literature, with most items
Figure 2.
Proposed research
model based on TAM
and SCT
Students’
intention to
accept AI-
based chatbot
adopted and a few adapted for relevance. The questionnaire, designed with five-point Likert
scales, measures latent variables and aligns with the study’s domain.
3.2 Validity of research instruments
The author engaged a psychometrician to ensure the questionnaire’s error-free construction
in the psychometric assessment (Kumar et al., 2021). Expert opinions validated research
instruments across multiple study dimensions. The questionnaire, aligned with a conceptual
benchmark, received feedback from 5 domain experts to assess depth and rationality of items.
Pilot testing (Mathews et al., 2023) involved 50 university students, refining the questionnaire
based on expert recommendations and respondent observations, confirming face validity
(Hock et al., 2023).
3.3 Data collection and respondents’profile
Table 1 presents student profiles selected for the study, detailing gender, age, academic year
and institution. The author distributed questionnaires via email, WhatsApp and social media
platforms from January to March 2023, receiving 420 responses. After data cleaning, 413
finalized responses were used for inferential analysis on the conceptual framework’s
measurement and structural models.
3.4 Common method bias test
The study uses self-administered questionnaires to gather responses on exogenous and
indigenous variables. Therefore, it assesses common method bias (CMB) using Harman’s
Single Factor test (Podsakoff and Organ, 1986;Elsayed, 2023). Fuller et al. (2016) and Saxena
et al. (2022) suggest suitability for statistical analysis if variance is <50%. Analysis indicates
a 31.21% variance in the data set, making the screened 413-sample data bias-free for valid
statistical analysis.
4. Data analysis
4.1 Analysis of measurement model
The study utilizes SPSS-25 for initial data cleaning and employs SmartPLS 4.0.9.2 for
descriptive and inferential analysis using PLS-SEM modeling. SPSS is preferred for data
cleaning, as it excels in identifying monotonous responses, handling missing values and
Profile dimensions Respondents’characteristics Participants (n5413)
Gender Male 216
Female 197
Age 21 79
22 145
23 189
Academic Year Second Year 80
Third Year 140
Fourth Year 193
Institutions University of Dhaka 103
University of Rajshahi 110
North South University 105
Bangladesh Army University of Science and Technology 95
Source(s): Authors’own creation
Table 1.
Respondents’
information
JARHE
testing data distribution (Roni and Djajadikerta, 2021). The research then utilizes PLS-SEM to
examine the reliability, validity and hypotheses of a complex model, acknowledging its utility
in differential psychology research with small sample sizes (Willaby et al., 2015).
4.1.1 Ensuring reliability and validity. 4.1.1.1 Goodness of fit measurement. Table 2
explores that the cleaned-up data set is perfectly fit for analysis with PLS-SEM since it meets
the threshold value. The analytical results show that the standardized root mean square
residual (SRMR) of the estimated model is 0.071 which is less than the threshold of 0.08
(Schuberth et al., 2022) and the normed fit index (NFI) is 0.953 which is very adjacent to 1
(Lohm€
oller, 1989).
4.1.1.2 Reliability and validity. Table 3 presents crucial metrics assessing the
measurement model’s robustness. It evaluates internal consistency, multicollinearity and
convergent validity through factor loadings, variance inflation factor (VIF), composite
reliability (CR), average variance extracted (AVE) and Cronbach’s alpha (CA). Bagozzi and Yi
(1988) suggested a factor loading range of 0.70–0.94 for effective latent variable
measurement. The finalized items, meeting this criterion, are displayed in Table 3. VIFs,
all below 5, confirm an absence of multicollinearity (Hair et al., 2013). CR and CA, gauging
model reliability, surpass the threshold of 0.70. AVE, indicating convergent validity, exceeds
0.50 (Hair et al., 2013,2020). Table 3 asserts that all criteria are met, affirming the model’s
reliability and convergent validity. These outcomes also affirm the measurement model’s
suitability for the structural model in the study.
4.1.1.3 Discriminant validity. This study assesses measurement model discriminant
validity using Fornell–Larcker criterion and Heterotrait-Monotrait ratio (HTMT) –Matrix.
Table 4 indicates latent variable AVEs exceed inter-correlations (Fornell and Larcker, 1981).
According to Henseler et al. (2014), all constructs in Table 5 meet the HTMT criterion’s 0.9
threshold, confirming discriminant validity.
4.2 Analysis of structural model
The present study applies a bootstrapping PLS-SEM calculation technique based on a 5000
resampling strategy to verify the direct relationship among latent variables (Hair et al., 2021;
Sarstedt et al., 2014;Alsyouf et al., 2023). Figure 3 shows the graphical representation of
structural equation modeling.
The study utilized partial least squares-structural equation modeling (PLS-SEM) to test six
hypotheses, as detailed in Table 6.Hypothesis 2 was the only one rejected, with PU not
significantly influencing IA (β50.009, Std Error 50.011, t 50.870, p50.384). Hypotheses 1,
3, 4, 5 and 6 were accepted. PEOU negatively impacted IA (β50.035, Std Error 50.010,
t53.508, p50.00). TSF, EWOM and ASE positively influenced IA, with βvalues of 0.514,
0.278 and 0.267, respectively. IA, in turn, positively impacted AU (β50.579, Std Error 50.035,
t516.70, p50.00). These results support the study’s structural model conclusions.
The study’s conclusions meet the criterion of R2 > 0.02 in order to confirm the data set’s
predictability of the dependent variable with the direct impact of independent variables
Saturated Model Estimated Model
SRMR 0.069 0.071
d_ULS 1.443 1.857
d_G 0.437 0.450
Chi-Square 765.436 781.416
NFI 0.863 0.953
Source(s): Authors’own creation
Table 2.
Goodness-of-fit index
Students’
intention to
accept AI-
based chatbot
Latent variable and sources Items FL VIF CR AVE CA(
α
)
Perceived ease of use (PEU)
Venkatesh et al. (2012),Ashfaq et al. (2020),Bilquise
et al. (2023),Li (2023)
PEU01 Delete N/A 0.805 0.580 0.639
PEU02 Delete N/A
PEU03 0.721 1.225
PEU04 0.781 1.340
PEOU05 0.781 1.235
Perceived usefulness (PU)
Venkatesh et al. (2012),Bilquise et al. (2023),
Li (2023)
PU01 Delete N/A 0.868 0.623 0.797
PU02 0.853 2.025
PU03 0.725 1.472
PU04 0.741 1.462
PU05 0.829 1.867
Time-saving feature (TSF)
Yeo et al. (2017),Hong et al. (2021)
TSF01 0.700 1.459 0.896 0.638 0.852
TSF02 0.736 1.568
TSF03 0.698 1.482
TSF04 0.914 2.23
TSF05 0.914 2.17
Electronic word-of-mouth (EWOM)
Choirisa et al. (2021)
EWOM01 0.784 1.416 0.881 0.650 0.822
EWOM02 0.853 2.497
EWOM03 0.847 2.623
EWOM04 Delete N/A
EWOM05 0.736 1.553
Academic self-efficacy (ASE)
Midgley et al. (2000),Bin-Nashwan et al. (2023)
ASE01 Delete N/A 0.876 0.638 0.812
ASE02 0.792 1.801
ASE03 0.820 1.852
ASE04 0.766 1.520
ASE05 0.818 1.635
Intention to adopt AI-based chatbot (IA)
Venkatesh et al. (2012),Bilquise et al. (2023),Roy et al.
(2022)
IA01 0.749 1.618 0.884 0.605 0.835
IA02 0.715 1.583
IA03 0.850 2.227
IA04 0.841 2.210
IA05 0.723 1.497
Actual use of AI-based chatbot (AU)
Li (2023)
AU01 Delete N/A 0.895 0.809 0.765
AU02 0.908 1.621
AU03 0.891 1.621
AU04 Delete N/A
Note(s): FL- Factor loading; AVE- Average variance extracted; CR- Composite reliability and CA-
Cronbach’s alpha
Source(s): Authors’own creation
ASE AU EWOM IA PEOU PU TSF
ASE 0.799
AU 0.507 0.900
EWOM 0.691 0.775 0.806
IA 0.932 0.579 0.863 0.778
PEOU 0.497 0.552 0.577 0.534 0.762
PU 0.716 0.466 0.686 0.757 0.498 0.789
TSF 0.940 0.558 0.806 0.978 0.529 0.745 0.799
Note(s): PEOU- Perceived ease of use; PU-Perceived usefulness; TSF- Time saving feature; EWOM-Electronic
word-of-mouth; ASE- Academic self-efficacy; IA- Intention to adopt AI-powered chatbot; AU- Actual use of
chatbot
Source(s): Authors’own creation
Table 3.
Convergent validity,
internal consistency
and multicollinearity
Table 4.
Fornell–Larcker
criterion
JARHE
(see Figure 4). This result shows that IA, as an independent variable, has predictability for AU
and that all five exogenous variables successfully predict IA, an endogenous variable.
Additionally, the PLS-SEM study results validate that the model can be predicted because
ASE AU EWOM IA PEOU PU TSF
ASE
AU 0.640
EWOM 0.811 0.776
IA 0.751 0.717 0.804
PEOU 0.682 0.783 0.800 0.726
PU 0.876 0.583 0.822 0.762 0.692
TSF 0.871 0.694 0.812 0.837 0.719 0.897
Note(s): PEOU- Perceived ease of use; PU-Perceived usefulness; TSF- Time saving feature; EWOM-Electronic
word-of-mouth; ASE- Academic self-efficacy; IA- Intention to adopt AI-powered chatbot; AU- Actual use of
chatbot
Source(s): Authors’own creation
Hypotheses Paths Beta (β) Std. Error t-value p-value Status
H1 PEOU→IA 0.035 0.010 3.508 0.000 Supported
H2 PU→IA 0.009 0.011 0.870 0.384 Not Supported
H3 TSF→IA 0.514 0.031 16.84 0.000 Supported
H4 EWOM→IA 0.278 0.015 18.74 0.000 Supported
H5 ASE→IA 0.267 0.027 9.73 0.000 Supported
H6 IA→AU 0.579 0.035 16.70 0.000 Supported
Note(s): PEOU- Perceived ease of use; PU-Perceived usefulness; TSF- Time saving feature; EWOM-Electronic
word-of-mouth; ASE- Academic self-efficacy; IA- Intention to adopt AI-powered chatbot; AU- Actual use of
chatbot
Source(s): Authors’own creation
Table 5.
Heterotrait-monotrait
ratio (HTMT) - Matrix
Figure 3.
Structural
equation model
Table 6.
Results for hypothesis
testing
Students’
intention to
accept AI-
based chatbot
they meet the f2 > 0.00 threshold (refer to Figure 4). Consequently, this indicates that the data
set and the model are a perfect match (Bin-Nashwan et al., 2023).
5. Discussion
The analysis part of the study reveals that PEOU significantly but negatively affects IA,
which offers strong support for hypothesis 1. The existing literature in this domain explored
the same outcomes where university students’PEOU has a negative influence on their IA (Liu
and Ma, 2023). This finding strongly suggests that university students’perception regarding
the easiness of using AI-powered digital chatbots does not determine their inner intention to
use chatbots. The result of the study also explores that PU insignificantly affects IA which
offers rejection for hypothesis 2. This finding does not completely match with existing
findings since the literature denotes that students’perception of the benefits of adopting AI-
powered chatbots significantly determines their intention to use them (Chocarro et al., 2023).
The analysis part of the study reveals that TSF positively and significantly affects IA, which
gives acceptance for hypothesis 3. The previous studies also explored the parallel result in
this field where TSF is regarded as a positive determinant of IA (Bin-Nashwan et al., 2023). On
the other hand, findings display that EWOM affects IA positively, which provides support for
hypothesis 4. The same findings were confirmed through previous studies in this domain
where the authors found that EWOM has a significant positive impact on students’IA (Bin-
Nashwan et al., 2023;Kusawat and Teerakapibal, 2022). The PLS-SEM result transparently
denotes that ASE as the model’s independent variable has a positive and statistically
significant impact on IA, which then offers acceptance for hypothesis 5. Meanwhile, this
finding is also supported by existing literature where the researchers obtained a positive and
significant influence of ASE on IA (Rudolph et al., 2023;Bin-Nashwan et al., 2023).
Finally, the statistical result from PLS-SEM informs us that IA has the strongest impact on
AU, which provides support for hypothesis 6. Existing literature (Ni and Cheung, 2023;Li,
2023;Rukhiran et al., 2023) verifies the result of this study by exploring IA as the significant
predictor to determine AU.
5.1 Theoretical implications
The present study contributes theoretically to this domain by comprehensively expanding
the existing literature. To the best of the author’s knowledge, widely applied TAM and SCT
Figure 4.
Results of structural
equation model
JARHE
research paradigms are integrated to create a new theoretical path for the very first time in
this study which will create an opportunity for future research in this domain. Furthermore,
this study might effectively enhance the readers’knowledge about the pros and cons of the
TAM and SCT to understand the university students’intention and actual use of AI-powered
digital chatbots such as ChatGPT. Previously, TAM and SCT were applied to investigate
behavioral responses separately. Therefore, this study develops a new foundation
undoubtedly based on the integration between TAM and SCT to investigate students’
behavioral intentions and consequences in the field of AI.
5.2 Managerial implications
The study’s results highlight key insights beneficial for policymakers, users and
communities. Notably, time-saving features in digital chatbots significantly influence
university students’adoption intentions. This insight aids developers in enhancing future
chatbot versions, emphasizing efficiency. Tailoring products to student preferences fosters
widespread acceptance globally. Additionally, the study assists universities in
understanding students’behavior influenced by cognitive and environmental factors.
Promoting awareness of AI-powered chatbot benefits increases adoption rates, such as with
ChatGPT. The findings deliver a crucial message to the student community, particularly in
higher education, potentially amplifying their inclination to use AI-powered chatbots based
on the study’s outcomes.
5.3 Knowledge implications
The findings significantly impact understanding technology adoption in education. It
highlights AI’s evolving role, emphasizing AI-powered chatbots as crucial student support
tools. The study provides insights into factors influencing student acceptance, aiding
educators and developers in enhancing user experience. Exploring students’intentions
contributes to the broader discourse on human-AI interaction, unveiling trust and acceptance
intricacies in educational technology. The findings guide educational institutions in seamlessly
integrating AI-powered chatbots, fostering a positive student-technology relationship.
References
Alam, M. (2022), “Supply chain management practices and organizational performance in
manufacturing Industry: SCM and organizational performance”,South Asian Journal of
Social Review (ISSN: 2958-2490), Vol. 1 No. 1, pp. 42-52, doi: 10.57044/sajsr.2022.1.1.2204.
Alsyouf, A., Lutfi, A., Alsubahi, N., Alhazmi, F.N., Al-Mugheed, K., Anshasi, R.J., Alharbi, N.I. and
Albugami,M.(2023),“The use of a Technology Acceptance Model (TAM) to predict patients’
usage of a personal health record system: the role of security, privacy, and usability”,
International Journal of Environmental Research and Public Health, Vol. 20 No. 2, p. 1347,
doi: 10.3390/ijerph20021347.
Awal, M.R., Hossain, M.S., Arzin, T.A., Sheikh, M.I. and Haque, M.E. (2023a), “Application of the SOBC
model to study customers’online purchase intentions in an emerging economy during
COVID-19: does gender matter?”,PSU Research Review. doi: 10.1108/prr-02-2022-0014.
Awal, M.R., Arzin, T.A., Islam, M.M. and Hasan, M.T. (2023b), “Understanding railway passengers’
E-ticketing usage intention in an emerging economic context: application of an extended
technology acceptance model”,Arab Gulf Journal of Scientific Research. doi: 10.1108/agjsr-12-
2022-0294.
Ashfaq, M., Yun, J., Yu, S. and Loureiro, S.M.C. (2020), “I, Chatbot: modeling the determinants of users’
satisfaction and continuance intention of AI-powered service agents”,Telematics and
Informatics, Vol. 54, 101473, doi: 10.1016/j.tele.2020.101473.
Students’
intention to
accept AI-
based chatbot
Astuti, E.S. (2023), “Collaborative attitudes and perceived usefulness to increase the intentions to use
internet banking”,International Journal of Electronic Finance, Vol. 12 No. 1, pp. 80-96, doi: 10.
1504/ijef.2023.10052677.
Bagozzi, R.P. and Yi, Y. (1988), “On the evaluation of structural equation models”,Journal of the
Academy of Marketing Science, Vol. 16, pp. 74-94.
Bailey, D.R., Almusharraf, N. and Almusharraf, A. (2022), “Video conferencing in the e-learning
context: explaining learning outcome with the technology acceptance model”,Education and
Information Technologies, Vol. 27 No. 6, pp. 7679-7698, doi: 10.1007/s10639-022-10949-1.
Bandura, A. (1986), Social Foundations of Thought and Action: A Social Cognitive Theory, Prentice-
Hall, Englewood Cliffs, NJ.
Bandura, A. (2023), Social Cognitive Theory: An Agentic Perspective on Human Nature, John Wiley
& Sons.
Bilquise, G., Ibrahim, S. and Salhieh, S.E.M. (2023), “Investigating student acceptance of an academic
advising chatbot in higher education institutions”,Education and Information Technologies,
pp. 1-26, doi: 10.1007/s10639-023-12076-x.
Bin-Nashwan, S.A., Sadallah, M. and Bouteraa, M. (2023), “Use of ChatGPT in academia: academic
integrity hangs in the balance”,Technology in Society, Vol. 75, 102370, doi: 10.1016/j.techsoc.
2023.102370.
Biswas, S.S. (2023), “Role of chat gpt in public health”,Annals of Biomedical Engineering, Vol. 51 No. 5,
pp. 868-869, doi: 10.1007/s10439-023-03172-7.
Chhibber, M. and Bhadauria, M. (2022), “ML enabled WhatsApp chatbot using IBM watson. Medicon
engineering themes”. doi: 10.55162/mcet.02.045.
Chocarro, R., Corti~
nas, M. and Marcos-Mat
as, G. (2023), “Teachers’attitudes towards chatbots in
education: a technology acceptance model approach considering the effect of social language,
bot proactiveness, and users’characteristics”,Educational Studies, Vol. 49 No. 2, pp. 295-313.
Choirisa, S.F., Purnamaningsih, P. and Alexandra, Y. (2021), “The effect of e-wom on destination
image and attitude towards to the visit intention in Komodo National Park, Indonesia”,Journal
of Tourism Destination and Attraction, Vol. 9 No. 1, pp. 49-60, doi: 10.35814/tourism.v9i1.1876.
Davis, F.D. (1989), “Perceived usefulness, perceived ease of use, and user acceptance of information
technology”,MIS Quarterly, Vol. 13 No. 3, pp. 319-340, doi: 10.2307/249008.
Elsayed, R.A.A. (2023), “The effect of investment in the brand value chain on profitability and market
value of the firm: lesson of success taken from Amazon”,Future Business Journal, Vol. 9 No. 1,
pp. 1-17, doi: 10.1186/s43093-023-00196-2.
Fishbein, M. and Ajzen, I. (1975), Belief, Attitude, Intention, and Behavior: an Introduction to Theory
and Research, Addison-Wesley, Reading, MA.
Fornell, C. and Larcker, D.F. (1981), “Evaluating structural equation models with unobservable
variables and measurement error”,Journal of Marketing Research, Vol. 18 No. 1, pp. 39-50,
doi: 10.2307/3151312.
Fuller, C.M., Simmering, M.J., Atinc, G., Atinc, Y. and Babin, B.J. (2016), “Common methods variance
detection in business research”,Journal of Business Research, Vol. 69 No. 8, pp. 3192-3198, doi:
10.1016/j.jbusres.2015.12.008.
Hair, J.F., Ringle, C.M. and Sarstedt, M. (2013), “Partial least squares structural equation modeling:
rigorous applications, better results and higher acceptance”,Long Range Planning, Vol. 46 Nos
1-2, pp. 1-12, doi: 10.1016/j.lrp.2013.01.001.
Hair, J.F. Jr, Howard, M.C. and Nitzl, C. (2020), “Assessing measurement model quality in PLS-SEM
using confirmatory composite analysis”,Journal of Business Research, Vol. 109, pp. 101-110,
doi: 10.1016/j.jbusres.2019.1.
Hair, J.F. Jr, Hult, G.T., Ringle, C.M. and Sarstedt, M. (2021), A Primer on Partial Least Squares
Structural Equation Modeling (PLS-SEM), Sage Publications, Thousand Oaks, CA, USA.
JARHE
Hajiyev, J. and Chang, C.T. (2017), “Gen y members’mobile banking adoption intention and actual use
in Azerbaijan and Turkey: the technology acceptance model and social cognitive theory
approach”,Journal of Internet Banking and Commerce, Vol. 22 No. S7, p. 1.
Hamidi, H. and Chavoshi, A. (2018), “Analysis of the essential factors for the adoption of mobile
learning in higher education: a case study of students of the University of Technology”,
Telematics and Informatics, Vol. 35 No. 4, pp. 1053-1070, doi: 10.1016/j.tele.2017.09.016.
Henseler, J., Dijkstra, T.K., Sarstedt, M., Ringle, C.M., Diamantopoulos, A., Straub, D.W., Ketchen, D.J.,
Hair, J.F., Hult, G.T.M. and Calantone, R.J. (2014), “Common beliefs and reality about PLS:
comments on r€
onkk€
o and evermann (2013)”,Organizational Research Methods, Vol. 17 No. 2,
pp. 182-209, doi: 10.1177/1094428114526928.
Hock, M., Tiringer, I., Ambrus, E., N
emeth, Z. and Farkas, B. (2023), “Validation and translation of the
Hungarian version of the Australian pelvic floor questionnaire (APFQ-H)”,International
Urogynecology Journal, Vol. 34 No. 6, pp. 1187-1194, doi: 10.1007/s00192-022-05322-2.
Holzinger, A., Keiblinger, K., Holub, P., Zatloukal, K. and M€
uller, H. (2023), “AI for life: trends in
artificial intelligence for biotechnology”,New Biotechnology, Vol. 74, pp. 16-24, doi: 10.1016/j.nbt.
2023.02.001.
Hong, C., Choi, H.H., Choi, E.K.C. and Joung, H.W.D. (2021), “Factors affecting customer intention to
use online food delivery services before and during the COVID-19 pandemic”,Journal of
Hospitality and Tourism Management, Vol. 48, pp. 509-518, doi: 10.1016/j.jhtm.2021.08.012.
Jiang, Y., Li, X., Luo, H., Yin, S. and Kaynak, O. (2022), “Quo vadis artificial intelligence?”,Discover
Artificial Intelligence, Vol. 2 No. 1, p. 4, doi: 10.1007/s44163-022-00022-8.
Kaczorowska-Spychalska, D. (2019), “Chatbots in marketing”,Management, Vol. 23 No. 1, pp. 251-270,
doi: 10.2478/manment-2019-0015.
Kumar, K.P., Sethi, B.K., Khare, S.P., Padsalge, M., Tripathi, A.P., Bhole, M. and Roy, D.S. (2021),
“Development of a symptoms-based thyroid risk questionnaire to screen for hypothyroidism in
India”,Thyroid Research and Practice, Vol. 18 No. 2, pp. 74-81, doi: 10.4103/trp.trp_5_22.
Kusawat, P. and Teerakapibal, S. (2022), “Cross-cultural electronic word-of-mouth: a systematic
literature review”,Spanish Journal of Marketing - ESIC.
Lee, Y., Lim, W. and Eng, H.S. (2023), “A systematic review of UTAUT2 constructs’analysis among
MSMEs in non-OECD countries”,Journal of Science and Technology Policy Management. doi: 10.
1108/jstpm-08-2022-0140.
Li, K. (2023), “Determinants of college students’actual use of AI-based systems: an extension of the
technology acceptance model”,Sustainability, Vol. 15 No. 6, p. 5221, doi: 10.3390/su15065221.
Li, H., Chen, Q., Zhong, Z., Gong, R. and Han, G. (2022), “E-word of mouth sentiment analysis for user
behavior studies”,Information Processing and Management, Vol. 59 No. 1, 102784, doi: 10.1016/
j.ipm.2021.102784.
Lin, C.C., Huang, A.Y. and Yang, S.J. (2023), “A review of ai-driven conversational chatbots
implementation methodologies and challenges (1999-2022)”,Sustainability, Vol. 15 No. 5, p. 4012,
doi: 10.3390/su15054012.
Liu, G. and Ma, C. (2023), “Measuring EFL learners’use of ChatGPT in informal digital learning of
English based on the technology acceptance model”,Innovation in Language Learning and
Teaching, pp. 1-14.
Liu, R., Menhas, R., Dai, J., Saqib, Z.A. and Peng, X. (2022), “Fitness apps, live streaming workout
classes, and virtual reality fitness for physical activity during the COVID-19 lockdown: an
empirical study”,Frontiers in Public Health, Vol. 10, 852311, doi: 10.3389/fpubh.2022.852311.
Lohm€
oller, J.-B. (1989), Latent Variable Path Modeling with Partial Least Squares, Physica, Heidelberg.
Martha, K. (2009), Factors Affecting Academic Performance of Undergraduate Students at Uganda
Christian University, Unpub. Master of Arts in Educational Management Dissertation,
Makerere University, Uganda.
Students’
intention to
accept AI-
based chatbot
Mathews, B., Meinck, F., Erskine, H.E., Tran, N., Lee, H., Kellard, K., Pacella, R., Scott, J.G., Finkelhor,
D., Higgins, D.J., Thomas, H.J. and Haslam, D.M. (2023), “Adaptation and validation of the
Juvenile Victimization Questionnaire-R2 for a national study of child maltreatment in
Australia”,Child Abuse and Neglect, Vol. 139, 106093, doi: 10.1016/j.chiabu.2023.106093.
Midgley, C., Maehr, M.L., Hruda, L.Z., Anderman, E., Anderman, L., Freeman, K.E. and Urdan, T.
(2000), Manual for the Patterns of Adaptive Learning Scales, University of Michigan, Ann
Arbor, pp. 734-763.
Na, S., Heo, S., Han, S., Shin, Y. and Roh, Y. (2022), “Acceptance model of artificial intelligence (AI)-
based technologies in construction firms: applying the Technology Acceptance Model (TAM) in
combination with the Technology–Organisation–Environment (TOE) framework”,Buildings,
Vol. 12 No. 2, p. 90, doi: 10.3390/buildings12020090.
Ng, K.S.P., Zhang, J., Wong, J.W.C. and Luo, K.K. (2023), “Internal factors, external factors and
behavioral intention toward food delivery apps (FDAs)”,British Food Journal, Vol. 125 No. 8,
pp. 2970-2987, doi: 10.1108/BFJ-07-2022-0586.
Ni, A. and Cheung, A. (2023), “Understanding secondary students’continuance intention to adopt AI-
powered intelligent tutoring system for English learning”,Education and Information
Technologies, Vol. 28 No. 3, pp. 3191-3216, doi: 10.1007/s10639-022-11305-z.
O’Connor, Y. and Mahony, C. (2023), “Exploring the impact of augmented reality on student academic
self-efficacy in higher education”,Computers in Human Behavior, Vol. 149, 107963, doi: 10.1016/
j.chb.2023.107963.
Podsakoff, P.M. and Organ, D.W. (1986), “Self-reports in organizational research: problems and
prospects”,Journal of Management, Vol. 12 No. 4, pp. 531-544, doi: 10.1177/
014920638601200408.
Raman, A., Thannimalai, R., Rathakrishnan, M. and Ismail, S.N. (2022), “Investigating the influence
of intrinsic motivation on behavioral intention and actual use of technology in moodle
platforms”,International Journal of Instruction, Vol. 15 No. 1, pp. 1003-1024, doi: 10.29333/iji.
2022.15157a.
Rana, N.P. and Dwivedi, Y.K. (2015), “Citizen’s adoption of an e-government system: validating
extended social cognitive theory”,SCT) Government Information Quarterly, Vol. 32 No. 2,
pp. 172-181, doi: 10.1016/j.giq.2015.02.002.
Roni, S. and Djajadikerta, H. (2021), “Preliminary data analysis: an analysis before the analysis. Data
analysis with SPSS for survey-based research”. doi: 10.1007/978-981-16-0193-4_3.
Roy, R., Babakerkhell, M.D., Mukherjee, S., Pal, D. and Funilkul, S. (2022), “Evaluating the intention
for the adoption of artificial intelligence-based robots in the university to educate the students”,
IEEE Access, Vol. 10, pp. 125666-125678, doi: 10.1109/access.2022.3225555.
Rudolph, J., Tan, S. and Tan, S. (2023), “ChatGPT: bullshit spewer or the end of traditional
assessments in higher education?”,Journal of Applied Learning and Teaching, Vol. 6 No. 1, doi:
10.37074/jalt.2023.6.1.9.
Rukhiran, M., Wong-In, S. and Netinant, P. (2023), “User acceptance factors related to biometric
recognition technologies of examination attendance in higher education: TAM model”,
Sustainability, Vol. 15 No. 4, p. 3092, doi: 10.3390/su15043092.
Sarstedt, M., Ringle, C.M. and Hair, J.F. (2014), “PLS-SEM: looking back and moving forward”,Long
Range Plan, Vol. 47 No. 3, pp. 132-137, doi: 10.1016/j.lrp.2014.02.008.
Saxena, M., Bagga, T., Gupta, S. and Kaushik, N. (2022), “Exploring common method variance in
analytics research in the Indian context: a comparative study with known techniques”,FIIB
Business Review, 23197145221099098, doi: 10.1177/23197145221099098.
Schuberth, F., Rademaker, M.E. and Henseler, J. (2022), “Assessing the overall fit of composite models
estimated by partial least squares path modeling”,European Journal of Marketing, Vol. 57
No. 6, pp. 1678-1702, doi: 10.1108/ejm-08-2020-0586.
JARHE
Tiwari, C.K., Bhat, M.A., Khan, S.T., Subramaniam, R. and Khan, M.A.I. (2023), “What drives students
toward ChatGPT? An investigation of the factors influencing adoption and usage of ChatGPT”,
Interactive Technology and Smart Education.
Venkatesh, V., Thong, J.Y. and Xu, X. (2012), “Consumer acceptance and use of information
technology: extending the unified theory of acceptance and use of technology”,MIS Quarterly,
Vol. 36 No. 1, pp. 157-178, doi: 10.2307/41410412.
Willaby, H., Costa, D., Burns, B., MacCann, C. and Roberts, R. (2015), “Testing complex models with
small sample sizes: a historical overview and empirical demonstration of what Partial Least
Squares (PLS) can offer differential psychology”,Personality and Individual Differences, Vol. 84,
pp. 73-78, doi: 10.1016/J.PAID.2014.09.008.
Yeo, V.C.S., Goh, S.K. and Rezaei, S. (2017), “Consumer experiences, attitude and behavioral intention
toward online food delivery (OFD) services”,Journal of Retailing and Consumer Services,
Vol. 35, pp. 150-162, doi: 10.1016/j.jretconser.2016.12.013.
Further reading
An, S., Eck, T. and Yim, H. (2023), “Understanding consumers’acceptance intention to use mobile food
delivery applications through an extended technology acceptance model”,Sustainability, Vol. 15
No. 1, p. 832, doi: 10.3390/su15010832.
Cope, B., Kalantzis, M. and Searsmith, D. (2021), “Artificial intelligence for education: knowledge and
its assessment in AI-enabled learning ecologies”,Educational Philosophy and Theory, Vol. 53
No. 12, pp. 1229-1245, doi: 10.1080/00131857.2020.1728732.
Huang, F., Teo, T. and Zhao, X. (2023), “Examining factors influencing Chinese ethnic minority
English teachers’technology adoption: an extension of the UTAUT model”,Computer Assisted
Language Learning, pp. 1-23, doi: 10.1080/09588221.2023.2239304.
Putri, G.A., Widagdo, A.K. and Setiawan, D. (2023), “Analysis of financial technology acceptance of
peer to peer lending (P2P lending) using extended technology acceptance model (TAM)”,
Journal of Open Innovation: Technology, Market, and Complexity, Vol. 9 No. 1, 100027, doi: 10.
1016/j.joitmc.2023.100027.
Corresponding author
Md. Rabiul Awal can be contacted at: rabiul.ru.mgt18@gmail.com
For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
Students’
intention to
accept AI-
based chatbot
