A THEORETICAL FRAMEWORK FOR AI-ENHANCED METACOGNITIVE REGULATION IN ACTIVE LEARNING ENVIRONMENTS

In recent years, there has been a core higher education debate that generative artificial intelligence (AI) is becoming increasingly prevalent, as AI tools can help to strengthen and weaken individuals' ability to take control over their own learning process. More recently, there have been signs of an actual type of cognitive offloading known as metacognitive laziness. The field still lacks an integrative theory outlining the conditions of increased effective metacognition regulation through the use of AI alongside metacognition when the two are utilized together. The purpose of this article is to go through the process of building a theoretical model that provides an explanation for the augmentation of metacognitive regulation processes in situations of active learning thanks to artificial intelligence systems. The design was qualified using a type of theory synthesis and conceptual and theory-building description, connecting to an integrative review of literature in first quartile journals and a congruence analysis of three theoretical bodies: self-regulated learning theory, explainable artificial intelligence frameworks, and problem-based learning. The outcome is the AI-Augmented Metacognitive Regulation model that focuses on the regulatory cycle, followed by three augmentation functions moderated by learner characteristics, a gradual control-transfer axis, and an active learning environment that provides genuine regulation demands. The model is expressed in a set of nine propositions that are testable and have implications for instructional design, system design, and assessment. The propositions it offers as a theory should be tested.