Page 647 - ISC PROCEEDINGS 21.4

TAM holds that perceived ease of use moderates perceived usefulness, and the two
constructs affect the attitude of the user towards technology, which in turn
predetermines the behavioral intention and the actual usage behaviour. The strength of
TAM was established in different technological settings, such as technology in education
and online learning platforms. Nevertheless, generative AI systems are not the same as
the conventional information systems because of their self-directed decision-making,
content-generating, and ethical considerations. As such, it is theoretically important to
extend TAM with other constructs to account for the complexities that lie behind the
adoption of AI in the educational context.
2.9.2. TAM Extension via Generative AI.
Three other constructs are added to the traditional TAM framework to overcome
the shortcomings of the latter; these are AI literacy, trust in AI, and social influence. Such
constructs are theoretically underpinned in regard to intelligent systems and the online
learning landscape.
3. Methodology
3.1. Research design
This study adopts a quantitative research approach to examine the factors
influencing students’ adoption of generative artificial intelligence (GenAI) technologies in
digital learning environments. A cross-sectional survey design was employed to collect
data from university students regarding their perceptions, attitudes, and behavioral
intentions toward generative AI tools.
The quantitative approach is appropriate because the study aims to test
hypothesized relationships among constructs derived from the extended Technology
Acceptance Model (TAM), including perceived usefulness, perceived ease of use, attitude
toward AI, social influence, trust, AI literacy, behavioral intention, and actual use.
Structural relationships among these variables were examined using multivariate
statistical techniques.
3.2. Population and sampling
The target population of this study consists of undergraduate and postgraduate
students enrolled in higher education institutions who have experience using digital
learning platforms and generative AI tools such as ChatGPT or similar systems.
A non-probability convenience sampling technique was employed due to
accessibility and feasibility considerations. Data were collected from students across
different academic disciplines to ensure diversity in educational background and
technological exposure.
A minimum sample size of 300 respondents was targeted to ensure adequate
statistical power for Structural Equation Modeling (SEM) analysis. According to SEM
guidelines, a sample size greater than 200 is considered sufficient for model estimation
and hypothesis testing.
3.3. Instrument development
Data were collected using a structured questionnaire developed based on validated
measurement scales from previous TAM and AI adoption studies. All items were adapted
to fit the context of generative artificial intelligence in higher education.
3.4. Data collection procedure
Data were collected through an online Google form survey distributed via university
email systems, student groups, and learning management systems. Participation was
voluntary, and respondents were informed about the purpose of the study.

646