The novel AI–Research Output (AI-RO) Model is developed and tested, which integrates the Technology Acceptance Model (TAM) and Socio-Technical Systems Theory to explain both the direct and conditional relationships between AI use and research output.
This study examines how generative artificial intelligence usage behavior (UG) influences the perceived research writing competence (RWC) of EFL university students in Ho Chi Minh City. Utilizing a sequential explanatory mixed-methods design, the inquiry integrates partial least squares structural equation modeling (PLS-SEM) with a thematic analysis of semi-structured interviews. This paper argues that while GenAI acts as a digital support tool, it also risks cognitive passivity that alters students' self-competence perceptions. The results show that the structural model explains 58.9% of the variance in perceived research writing competence (R2 = 0.589). The frequency of GenAI interaction directly supports learners' writing confidence (β = 0.265, p < 0.001), while active academic discourse socialization (AD) serves as a constructive mediating channel (β = 0.133, p < 0.001) that enables students to internalize discipline-specific rhetorical conventions. Conversely, extensive technological interaction induces a state of cognitive dependency (CD) (β = 0.641, p < 0.001), yielding an inflated perceived competence score (β = 0.233, p < 0.001). This dual pathway reveals a paradoxical educational phenomenon whereby algorithmic text-processing efficiency is frequently mistaken for independent scholarly reasoning. Since the indirect effect of CD is larger than that of AD, these outcomes highlight the risk of technological over-reliance. In practical terms, these findings suggest that higher education institutions should adapt assessment paradigms and implement structured metacognitive training.
Quoc Hoang, Minh Nam Anh Nguyen· Technium Social Sciences Jou...· 0 citations
Artificial Intelligence (AI) assisted writing tools, such as ChatGPT have arisen rapidly and created a socio-technical phenomenon of fast technology adoption in higher education. Traditional models focus on early adoption but understanding the behavioral pathway that accounts for intensive AI overuse is critical. To summarize, this study enriched the Technology Acceptance Model (TAM) by developing a sequential mediation model: Perceptions (PU/PEOU) → Satisfaction → Habit → AI Overuse and investigated boundary conditions of this process regarding Confucian Cultural Values as moderator. By conducting the survey mixed-mode, from December 2024 to February 2025 data has been collected using a systematic random sampling of Chinese educators at six universities in Chongqing through both individual intercepting and online distribution. Direct, mediation and moderation effects were tested based on Partial Least Squares Structural Equation Modeling (PLS-SEM). Results confirm partial mediation: utility and ease perceptions drive user satisfaction, to the extent that it spurs behavioral automaticity (habit) behind high-intensity use the study characterizes AI Overuse as a relatively high force and goal-oriented behavioral outcome of technology acceleration, separating this from clinical addiction. As for the ability to moderate, Confucian Values firmly bolster the impact of Perceived Usefulness upon Satisfaction according to the power placed on collective welfare. Interestingly, these cultural effects are moderated when the variables Perceived Ease of Use (a cognitive barrier) and Habit (a psychological boundary) are applied. Theoretically, the current research offers a new lens through which accelerated technology adoption can be analysed and guides tailored digital sustainability interventions in Confucian models of education.
Jingyi Li, Lianyun Huang, F. Furuoka· Frontiers in Psychology· 0 citations
Background: The rapid spread of AI tools in education has moved the conversation from simply adopting technology to examining how it actually transforms what teachers do in the classroom. Yet most studies have tackled this question by looking at one variable at a time, leaving the structural pathways that connect readiness to real instructional outcomes largely unexplored. Aims: This study aims to investigate the structural relationships between teachers' technological readiness, specifically digital literacy, self-efficacy, and attitudes toward AI, and their actual instructional performance, while exploring the mediating role of AI competence. Methods: Survey responses from 220 teachers drawn from across the Indonesian archipelago were examined using Partial Least Squares Structural Equation Modeling (PLS-SEM) in SmartPLS 4, proceeding through a two-stage measurement and structural assessment. Results: Teacher Digital Literacy (TDL), Attitude Towards Using AI Tools (ATT), and Self-Efficacy with Technology (SET) each proved to be significant predictors of TAC (p < 0.001). Of all the variables tested, TAC turned out to be the strongest direct predictor of TP (β = 0.540). A more striking finding was that ATT carried no meaningful direct influence on TP (β = 0.009, p = 0.870); instead, its entire effect flowed through TAC, a pattern that satisfies the conditions for full mediation. Multi-Group Analysis further confirmed that these relationships held regardless of how long a teacher had been in the classroom.Conclusion: What these results collectively suggest is that enthusiasm alone cannot sustain instructional quality in the AI era; it is technical mastery in AI Competence that converts a teacher’s positive outlook into measurable performance gains.
Muhammad Ridzik Varedho, H. Harjono, Erisa Kurniati· FINGER : Jurnal Ilmiah Tekno...· 0 citations
This qualitative study investigates researchers' perceptions of ChatGPT's influence on research creativity, efficiency, and scholarly integrity. A cross-sectional survey with six open-ended questions was administered to 112 participants across academic levels (2 post-doctoral, 20 PhD, 26 Master's, 64 diploma holders) within education and psychology. Data was analyzed using triangulated methods: SWOT analysis, thematic analysis, and systematic open coding. Findings indicate ChatGPT is perceived as a dual-edged tool that enhances productivity while posing cognitive and ethical risks. Ethical considerations are central, with participants emphasizing the need for structured guidelines. The human factor remains decisive—AI's benefits depend on researchers' methodological awareness and ethical engagement. Results suggest AI functions optimally as a cognitive scaffold, with its impact contingent upon use patterns and user experience. The findings carry implications for developing evidence-based policies and training for responsible AI integration in academic research.
M. Moussa, Medhat Mohamed Saleh, O. Al-Adamat· Journal of Digital Education...· 0 citations
With the advancement of artificial intelligence (AI) technology, AI tools like ChatGPT have gradually become a part of the life of students in universities. Previous research has mainly addressed the functional aspects of AI, such as information retrieval and academic support, but the socio-emotional and personalized aspects of AI, and how these shape users’ continued usage behaviour. Therefore, this study aims to examine the effects of academic support, emotional support, and perceived personalisation on students’ continuance intention toward ChatGPT, with perceived trust acting as a mediating mechanism. This research grounded by the Expectation-Confirmation Theory and Information Systems Continuance Model by incorporating two additional variables: perceived personalisation and academic support and emotional support as antecedents of students’ continuance intention toward ChatGPT. This quantitative study used an online survey to collect data from 263 university students. The proposed model was analysed using SmartPLS-SEM to examine both direct and indirect relationships, with perceived trust specified as a mediating variable. The structural model indicates that academic support and emotional support have significant direct effects on continuance intention. Although perceived personalisation positively influences user trust, its direct effect on continuance intention is not significant. Instead, perceived trust fully mediates this relationship, indicating that personalisation alone does not directly drive continued use unless it first builds user trust. These findings highlight the critical role of trust as a psychological mechanism through which AI personalisation translates into sustained user engagement. The study contributes to the literature on human–AI interaction by integrating cognitive, emotional, and personalisation dimensions within a unified continuance framework. Practically, the results suggest that AI developers should prioritise trust-building mechanisms when designing personalised and emotionally supportive systems to ensure long-term user adoption, particularly in educational contexts.
P. Muthurajan, Hashima Mohaini Mohammad, Nur Afni Halil et al.· International Journal of Mod...· 0 citations
Generative AI (GenAI) is now common in university project work, yet previous studies often examine students’ trust, creativity, overload, or engagement separately. This leaves a key gap: how students regulate GenAI across a full project workflow. This exploratory study addresses that gap by examining a four-process “
AI-mediated Self-Regulated Learning
” (
AI-SRL
) cycle: (1) evaluating and selecting GenAI suggestions, (2) experiencing shifts in creative agency, (3) managing overload through filtering and summarizing, and (4) monitoring time and energy to stop or continue working with GenAI. We conducted a two-course basic qualitative design study with 97 undergraduate and graduate students. Data came from an open-ended questionnaire aligned to the four processes. We used inductive content analysis with a shared codebook, reliability checks, and cross-level comparisons. Findings show that students use combinations of strategies across the
AI-SRL
cycle. They exercise agency through goal alignment, revision, and verification, with graduates reporting stronger cross-checking and source-based justification. Creativity was described as a conditional outcome: it increased when GenAI widened ideas but declined when it replaced personal exploration. Overload was managed through targeted prompts and structured outputs, again more common among graduates. Most students did not lose track of time; they used clear stopping cues such as fatigue, repetition, or satisfaction. Together, results reveal two distinct metacognitive regulation styles: “
Exploratory-Simplification
” and “
Systematic-Methodical
”.
Maria A. Perifanou, Anastasios A. Economides· Journal of educational compu...· 0 citations
A new method for surgically removing training examples from a model reveals that as datasets grow, the link between what a model learns and what it produces dissolves.
