Implementing AI-Supported PBL in Higher Education: A Literature Review

The integration of Artificial Intelligence (AI) into Problem-Based Learning (PBL) has gained attention in higher education, yet a comprehensive synthesis of empirical evidence remains limited. This study systematically reviews literature on AI-supported PBL implementation in higher education, focusing on AI integration models, effectiveness on learning outcomes, and implementation challenges. A systematic review following PRISMA guidelines was conducted using open-access databases, covering publications from 2015 to 2025. After a four-stage selection process, empirical studies met the inclusion criteria and were synthesized narratively. The findings reveal three AI integration models: AI for problem identification and scenario generation, AI for self-directed learning and resource recommendation, and AI for synthesis, reflection, and feedback. The effectiveness of AI-PBL on cognitive outcomes was moderate to large, particularly in critical thinking and complex problem-solving, while affective outcomes were moderated by students' self-regulated learning. Implementation challenges across technical, pedagogical, and ethical dimensions were identified. This study concludes that AI holds promise for enhancing PBL, but success depends on pedagogical design, infrastructure readiness, and balancing AI assistance with students' critical thinking autonomy.