Artificial intelligence applications in education: A systematic review and meta-analysis on students' academic achievement and motivation

This doctoral dissertation aims to systematically examine the effects of artificial intelligence (AI) applications in education on students' academic achievement and learning motivation, as well as to explore how these effects vary across contextual variables such as educational level, type of application, subject area, and cultural context. The study was conducted within the framework of a systematic review and meta-analysis, grounded in an evidence-based research approach. A total of 181 experimental and quasi-experimental studies published between 2003 and 2024 were collected following PRİZMA guidelines from databases including YÖK Thesis Center, ProQuest, Web of Science, ScienceDirect, EBSCO, and Scopus. The meta-analysis results revealed a strong positive effect of AI-supported instructional practices on students' academic achievement (Cohen's d = 0.83; 95% CI [0.70, 0.96]) and a moderate but significant effect on learning motivation (d = 0.77; p < .01). The heterogeneity test (Q = 2236.75; 207.032 p < .01) indicated substantial contextual and methodological diversity among the studies. Moderator analyses showed that effect sizes significantly varied by educational level, application type, subject area, intervention duration, and cultural background. The highest effects were observed in the fields of social sciences (d = 1.51) and visual arts (d = 1.26), while lower effects were reported in basic sciences (d = 0.36). Intelligent Tutoring Systems (d = 1.52) and deep learning–based applications were identified as the most effective in enhancing personalized learning and self-regulation skills. Overall, the findings indicate that AI functions not merely as a cognitive enhancement tool but also as a pedagogical catalyst that strengthens students' affective and motivational dimensions. However, the observed variations highlight that AI's educational impacts are contingent upon contextual factors and pedagogical design features. This underscores the need to approach AI integration in education from a holistic perspective, encompassing pedagogical, ethical, and methodological dimensions rather than focusing solely on technological adaptation. This study contributes both theoretically and practically to the development of educational policies, instructional design, and digital learning strategies by providing comprehensive evidence on the effectiveness of AI-enhanced learning practices.