Shaumya Shriya
Magadh University, Bodhgaya, Bihar
Shashidhar Gupta
S. N.Sinha College, Jehanabad
Pages:144-154
The fast adoption of the concept of Artificial Intelligence (AI) in learning institutions has
encouraged the emerging research focus on the psychological effects of AI on students.
The present meta-analytic review has a systematic and organized investigation of the
impacts of AI-based interventions on three constructs that are significant, including
learner autonomy, self-efficacy, and psychological well-being. The search was
conducted as per the Preferred Reporting Items of Systematic Reviews and Meta-
Analyses (PRISMA) guidelines, whereby after selecting six large databases, 1,246
studies that were published in the year 2018 to 2025 were initially found. The final
number of 38 studies to be quantitatively synthesized was after the strict process of
eligibility screening, elimination of duplicates and quality assessment. The overall effect
sizes were calculated based on a random-effects model to take into consideration the
heterogeneity of the various educational settings, modes of AI and demographics of
the participants. The results point to a moderate positive impact of AI-based interventions
on self-efficacy and a small, but significant positive impact on learner autonomy.
Nevertheless, the correlation between integration of AI and psychological well-being
was subtle, with the level of human supervision and a sense of learner control. Subgroup
analysis also suggests that adaptive AI systems that maintain learner agency achieve
well-being results that are more powerful compared to all-automated solutions. The
review adds to a body of evidence that is coherent and provides practical suggestions
to craft psychologically informed AI-enhanced learning environments