Jul 2026· Technium Social Sciences Journal· Vol 85, pp. 240-264· 0 citations
TL;DR
It is argued that while GenAI acts as a digital support tool, it also risks cognitive passivity that alters students' self-competence perceptions, and higher education institutions should adapt assessment paradigms and implement structured metacognitive training.
Abstract
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.
This study used a convergent parallel mixed-methods design to examine at how generative Artificial Intelligence (AI) is reshaping educational practices, however there is limited evidence regarding its impact on students’ learning outcomes, especially in developing countries. This study further examined the relationships among AI-related skills, perceived usefulness, self-efficacy, motivation, and academic performance, as well as students’ experiences with generative AI in learning environments. Applying a survey design, this research collected survey data from two hundred ninety-nine ( N = 299) students in Nepal, Indonesia, and Brazil. This study also conducted semi-structured interviews with fourteen ( N = 14) interviewees. Quantitative results indicated low levels of AI literacy, self-efficacy, perceived usefulness, and motivation, all averaging means below midpoint scores of five point Likert scale of survey variables. Regression analyses revealed weak correlations between those independent variables and reported academic enhancement, suggesting that students often lack the confidence and skills required to integrate AI tools effectively into their learning activities.
Conversely, qualitative results highlighted significant advantages of generative AI, such as improved learning efficiency, enhanced communication, and greater problem-solving capabilities. Interviewees noted that AI tools simplified complex concepts and saved time, despite receiving limited formal training. The integration of quantitative and qualitative results exposed a perception practice gap: while students benefited from generative AI, they generally underestimated their abilities and lacked structured guidance. The study concludes that successful AI integration in education required more than mere access to technology; it requires fostering AI literacy, self-efficacy, motivation, and ethical awareness through dedicated pedagogical support.
Basanta Prasad Adhikari, Suyantiningsih, Ariyawan Agung Nugroho et al.· OCEM Journal of Management,...· 0 citations
Feedback is a core mechanism of formative assessment, yet its effect on learning depends on how students conceive of and act upon it. This study tests an AI-enhanced student conceptions of feedback model in English as a Foreign Language (EFL) writing, relating six feedback conceptions to academic self-efficacy and self-regulation. Survey responses from 300 undergraduate students at a state university in Indonesia were analysed with partial least squares structural equation modelling (PLS-SEM) using 5,000 bootstrap subsamples. The measurement model met every reliability and validity criterion: outer loadings ranged from 0.743 to 0.888, composite reliability from 0.857 to 0.931 and average variance extracted from 0.609 to 0.750, while the Fornell-Larcker criterion, cross-loadings and HTMT ratios all confirmed discriminant validity. All seven structural paths were significant. Active use of AI feedback was the strongest antecedent of self-efficacy, followed by enjoyment, whereas ignoring feedback exerted a significant negative effect. Self-efficacy in turn strongly predicted self-regulation and accounted for 43.5% of its variance. The results indicate that what students do with feedback matters more than the source from which it originates, and that self-efficacy is the pivotal mechanism translating feedback experience into self-regulated learning.
Virgiawan Adi Kristianto, Sucipto, W. Sumbodo et al.· MATEC Web of Conferences· 0 citations
This research aims to investigate how generative AI literacy, self-efficacy, attitude, interest and dependence interact to influence academic work completion among university students in Ghana. It also seeks to identify the psychological pathways through which AI competence is associated with students' behaviour in the context of increasing generative AI adoption.
This research used a quantitative cross-sectional study to collect the data from 466 undergraduate students at KNUST. Partial least squares structural equation modelling (PLS-SEM) was used to analyse the data among five constructs.
The results reveal a progressive behavioural pathway in which generative AI literacy positively predicts students' attitudes toward AI, which in turn strengthens their interest. This heightened interest significantly enhances students' self-efficacy in using AI, ultimately leading to dependence-induced task completion. Notably, self-efficacy emerged as the strongest predictor of task completion, underscoring both the empowering potential of AI use and the risk of increasing reliance on AI for academic work.
The cross-sectional design of this study limits its ability to interpret causality between the variables examined. Again, self-report measures are subject to common-method bias. Since the sample involved only KNUST undergraduate students, generalisation of the results cannot be assumed for other academic institutions. Finally, Partial Least Squares Structural Modelling assumes a linear relationship amongst all factors, thus may miss a nonlinear relationship that occurs within students' behaviours.
The findings provide actionable guidance for universities and policymakers seeking to advance SDG 4 (Quality Education) through responsible generative AI integration. Higher education institutions should embed structured AI literacy and ethical-use training within curricula to strengthen student self-efficacy while preventing unhealthy dependence. Assessment practices should be redesigned to emphasise critical thinking, creativity and reflective engagement rather than automated task completion. At the policy level, national higher education authorities are encouraged to develop AI-use guidelines that promote equity, academic integrity and learner autonomy. These practices support Emerald's impact agenda by translating empirical evidence into scalable educational interventions that enhance learning quality, student welfare and sustainable digital transformation in higher education.
The study underscores the need for institutional and national policies that guide responsible generative AI use in higher education while protecting student welfare. As AI literacy and self-efficacy increase, so does the risk of excessive dependence, with potential consequences for independent thinking, academic integrity and long-term cognitive development. Universities, particularly in developing contexts such as Ghana, must establish clear AI governance frameworks, embed ethical AI literacy into curricula and redesign assessments to prioritise critical engagement over automated outputs. Policy interventions should also ensure equitable access to AI training and support systems that promote student autonomy, well-being and sustainable learning practices in AI-mediated academic environments.
This study provides an understanding of GenAI use in African higher education by integrating AI literacy, self-efficacy, attitude, interest and dependence into one behavioural model that has been clearly tested in the Ghanaian context. The study provides new insights into how psychological factors and usage-based patterns shape students' dependence on Generative AI for class task completion, providing great insight for teachers who seek to continuously integrate AI into their work. The findings also support curriculum design and policy development by identifying the competencies and behavioural risks that must be addressed to guide effective AI use within universities.
Jacinta Mandy Baah, Eric Nana Ayeh Ntow, Jesse Ohene Boakye et al.· Journal of Applied Research...· 0 citations
The rapid integration of generative artificial intelligence (GenAI) in higher education has recontextualized cognitive development, particularly regarding Bloom’s higher-order thinking skills (HOTS). While GenAI's potential to enhance learning is widely recognized, the mechanisms through which GenAI-assisted HOTS influence academic achievement remain underexplored. This study addresses this gap by examining the mediating role of multidimensional student engagement (behavioral, emotional, and cognitive) in the relationship between GenAI-supported HOTS and academic achievement. Employing a quantitative, cross-sectional design, data were collected from 504 higher education students in Islamabad. Structural equation modeling (SEM) and bootstrap mediation analyses were utilized to test the hypothesized framework. Results indicated that GenAI-assisted HOTS significantly predicted both student engagement (β = .62, p < .001) and academic achievement (β = .28, p < .001). Crucially, student engagement partially mediated the relationship between GenAI-HOTS and academic achievement, accounting for 50.9% of the total effect (β_indirect = .29, p < .001), with cognitive engagement emerging as the strongest mediating pathway. Multi-group analysis confirmed the structural invariance of this model across academic levels and disciplines. These findings demonstrate that GenAI's impact on academic achievement is significantly channeled through enhanced student engagement rather than occurring solely through direct cognitive offloading. The study provides empirical evidence for shifting pedagogical strategies from passive AI consumption to structured, engagement-driven AI integration that explicitly targets higher-order cognitive processes.
Sidra Saeed, Hina Shaukat, Muhammad Minhass Baloch et al.· Journal of Global Social Tra...· 0 citations
Introduction Grounded in the Job Demands-Resources (JD-R) model, this study investigates the relationships between teacher authenticity, professional identity, psychological capital (PsyCap), and work engagement within mainland China’s educational system, focusing on K-12 teachers (N = 627) navigating high-stakes testing environments. Methods Employing an explanatory sequential mixed-methods design, the quantitative phase utilized structural equation modeling (SEM) to test a hypothesized mediation model. A subsequent qualitative phase (N = 26) involved thematic analysis to corroborate findings and holistic critical case analysis to explain quantitative outliers or “deviant” cases. Results Quantitative findings demonstrated that authenticity and professional identity significantly predict work engagement, both directly and indirectly through PsyCap, which emerged as a robust mediator (authenticity total effect β = 0.390, p < 0.001; professional identity total effect β = 0.485, p < 0.001). While thematic analysis corroborated these patterns, the critical case analysis revealed that these positive relationships are contingent upon organizational alignment and career stage. Discussion These results suggest that while authenticity and identity are vital personal resources, their efficacy in driving engagement is moderated by the institutional context. The study offers nuanced implications for fostering teacher well-being within high-stakes educational environments.
Guided by Sociocultural Theory, this study explores micro-level peer scaffolding strategies used by symmetrical EFL university students across pre-, while-, and post-writing stages. Using a mixed-methods design, data from audio recordings, interviews, and questionnaires (N=150) were analyzed via an expanded coding framework. Results show scaffolding is stage-sensitive: conceptual strategies (e.g., Questioning–Clarifying) dominated pre-writing; strategic moves (e.g., Stating–Modeling) guided drafting; and evaluative strategies (e.g., Suggesting–Rewording) supported revision. Affective scaffolding, though less frequent, underpinned a collaborative climate. A key finding was the gap between observed frequent questioning and learners' lower self-reported use, suggesting culturally mediated implicit mediation. The study offers a stage-based scaffolding model, demonstrating how reciprocal peer dialogue within a shared ZPD transforms writing from an individual task into a co-constructed process, with implications for fostering autonomy and competence in collaborative EFL writing instruction.
Nguyen Van Huong· Theory and Practice in Langu...· 0 citations