Human–GenAI collaboration: a framework for engagement, self-regulation and ethical learning outcomes

Purpose Generative Artificial Intelligence is reshaping higher education by enabling new forms of collaboration between learners and intelligent systems. Drawing on Hybrid Intelligence and Self-Regulated Learning theories, this study proposes and empirically tests a Hybrid–Self-Regulation Model to explain how human–artificial intelligence collaboration influences engagement, self-regulation and learning outcomes. Specifically, the study examines how perceived collaborative artificial intelligence support, trust in artificial intelligence collaboration and artificial intelligence-enabled adaptive personalization drive collaborative engagement (CE), which in turn enhances self-monitoring (SM) and metacognitive reflection (MR). Design/methodology/approach Using data from 307 students across six countries, the study examines how Perceived Collaborative AI Support (PCAS), Trust in AI Collaboration (TAIC) and AI-Enabled Adaptive Personalization (AIAP) drive CE, which in turn enhances SM and MR. Findings The findings reveal that perceived collaborative artificial intelligence support, trust in artificial intelligence collaboration and artificial intelligence-enabled adaptive personalization significantly increase CE and CE strongly predicts SM and MR. MR further positively influences academic achievement, creativity and innovation and responsible and ethical problem-solving. Serial mediation analyses confirm that the impact of perceived collaborative artificial intelligence support and artificial intelligence-enabled adaptive personalization on learning outcomes is transmitted through CE, SM and MR. Originality/value This study contributes theoretical clarity on human–artificial intelligence co-learning mechanisms and offers practical guidance for designing artificial intelligence-enhanced, ethically grounded learning ecosystems in higher education.