AI-Powered Game-Based Learning for Project Management Education

Bridging the gap between theory and practice remains a core challenge in project management education, as students often struggle to transfer classroom knowledge into complex, uncertain, and ethically charged environments. As serious games increasingly integrate artificial intelligence (AI), new opportunities arise for scalable, adaptive, and ethically informed learning. This paper introduces and evaluates an AI-enhanced Game-Based Learning (AI-GBL) framework that positions AI as a pedagogical co-orchestrator—modulating narrative and challenge—rather than a content add-on. It embeds serious game mechanics through two interactive modalities: AI-driven stakeholder interviews and collaborative crisis scenarios requiring students to make decisions, weigh trade-offs, and reflect on consequences. Structured debriefs and ethically ambiguous dilemmas reinforced reflection, judgment, and leadership identity. Within this framework, the lecturer served as a game master, shaping the environment, guiding discussion, and ensuring ethical engagement, while AI dynamically adapted gameplay. Grounded in Kolb's experiential learning cycle and scaffolded by the 5E model, the design also drew on Self-Determination Theory and Flow to support autonomy, competence, and engagement. A mixed-methods, quasi-experimental pretest–posttest design was applied in live classrooms. Undergraduates (n = 113) were grouped by tutorial exposure (experimental: ≥3, n = 66; control: ≤2, n = 47), while a postgraduate cohort (n = 47) contributed qualitative reflections. Students in the AI-GBL group achieved significantly higher posttest scores (means: 66.64 vs. 48.00). ANCOVA controlling for pretest confirmed a large effect (partial η² = .799), though this should be interpreted cautiously due to the non-randomised design. Qualitative reflections revealed growth in ethical reasoning, collaborative identity, and adaptive decision-making, alongside signs of over-reliance on AI—highlighting the need for scaffolding in ethical AI literacy. Findings suggest that AI-GBL can enhance domain-specific performance and the metacognitive and affective dimensions of adaptive expertise—ethical sensitivity, reflective judgment, and leadership identity. Three design principles emerge: (1) AI should function as a co-orchestrator, modulating challenge and feedback in real time; (2) open-ended, ethically complex scenarios deepen reflection beyond procedural accuracy; and (3) critical scaffolding is essential to prevent passive reliance on AI-generated outputs. This study builds on prior conceptual work published in Project Leadership and Society by providing empirical evidence of AI-GBL's effectiveness, advancing design principles for AI-integrated serious games, and identifying directions for future research, including longitudinal tracking, replication, and cross-domain application.