AI-supported learning and student outcomes: a systematic review informed by self-determination theory (SDT)

Abstract

Artificial intelligence (AI) is increasingly integrated into educational contexts to support personalised learning and adaptive instruction. This systematic review synthesises empirical studies published between 2018 and 2025 to examine the relationships between AI-supported learning and students’ motivation, engagement, and academic achievement through the lens of self-determination theory (SDT). The review draws on evidence from diverse educational contexts, ranging across the school-age range to higher education, and across multiple countries. Findings indicate that AI interventions, including generative AI tools (e.g. ChatGPT, Co-pilot), intelligent tutoring systems, and gamified learning environments, are generally associated with enhanced motivation and engagement by supporting autonomy, competence, and personalised learning experiences. Improvements in academic outcomes, particularly in language learning, writing quality, and problem-solving, are most evident when AI is pedagogically aligned and implemented with adaptive scaffolding. Consistent with SDT, support for autonomy and competence is well established across studies, while support for relatedness is less consistently addressed. However, excessive or unguided AI use may reduce opportunities for social interaction, potentially limiting engagement. Overall, AI-supported learning appears most effective when it is pedagogically grounded, context-sensitive, and carefully integrated into instructional practice through teacher–AI–learner interaction models.