Generative AI as a Metacognitive Co-Regulator in Training and Development: An Integrated Bibliometric and Systematic Review

Silvi Novrian Yulandari (State University of Surabaya)
Fida Rachmadiarti (State University of Surabaya)
Mita Anggaryani (State University of Surabaya)
Alif Syaiful Adam (National Taiwan University of Science and Technology)

Prisma Sains: Jurnal Pengkajian Ilmu dan Pembelajaran Matematika dan IPA IKIP Mataram
Vol. 14 No. 3: July 2026

The rapid integration of Generative Artificial Intelligence (GenAI) into organisational learning environments has reshaped how employees plan, monitor, and evaluate their learning. However, the role of GenAI as a metacognitive co-regulator in Training and Development (T&D) remains conceptually fragmented across research on metacognition, AI-supported learning, and global talent development. Drawing on an integrated bibliometric analysis and systematic review approach, this study examines the intellectual structure, thematic evolution, and conceptual intersections of these three domains. A structured search of Scopus and Web of Science, supplemented by Google Scholar, identified 196 records published between 2000 and 2025. After duplicate removal and screening following an adapted PRISMA protocol, 139 records were retained for bibliometric mapping using Biblioshiny, and 139 full-text studies were included in the qualitative synthesis. Using bibliometric mapping techniques, the study identifies dominant research trends, functional roles of GenAI, and underexplored thematic gaps. The analysis reveals three dominant thematic clusters, with metacognition and self-regulated learning occupying the motor-theme quadrant and GenAI-related themes moving rapidly toward conceptual centrality, while global talent development remains peripheral. A complementary systematic qualitative synthesis further clarifies how GenAI supports metacognitive regulation across planning, monitoring, and evaluation phases. The integrated evidence reveals a conceptual shift from automation-oriented AI applications toward learner-centred metacognitive scaffolding, while highlighting the limited integration of AI-supported metacognition within global talent development frameworks. Based on these findings, the study proposes the GenAI–Metacognitive Workforce Development (GMWD) model, which conceptualises GenAI as a phase-sensitive metacognitive co-regulator that preserves learner agency while fostering adaptive performance, self-regulated professional learning, and global competence. The model provides a theoretically grounded framework for advancing research and practice in AI-enhanced training and development (T&D), offering practical implications for designing reflective, adaptive, and globally oriented learning environments for the workforce.