Abstract:

This paper presents an extended theoretical development of the STAC framework (Study–Think–Act–Connect), an instructional model designed for higher education in AI-mediated learning environments. The objective of the contribution is to construct an integrated theoretical foundation that combines classical learning theories with recent research on AI-supported scaffolding, addressing the increasing cognitive complexity of contemporary teaching and learning contexts.

The study adopts a bibliographic and theoretical–conceptual methodology organised into three sequential phases. First, a structured identification of theoretical and empirical sources is conducted, reviewing contributions from Ausubel’s meaningful learning, Vygotsky’s Zone of Proximal Development, Bruner’s scaffolding, and Kolb’s experiential learning cycle, together with recent empirical studies on AI-mediated instructional support. This includes research on adaptive scaffolding, cognitive load regulation, and AI-based metacognitive prompting, which indicates that generative AI can be used not to eliminate cognitive load, but to regulate it after intentional exposure to complex content, allowing learners to engage with higher levels of difficulty in shorter timeframes before knowledge gaps are closed through structured instructor-led scaffolding. Second, an integrative synthesis of the literature identifies shared cognitive mechanisms across theoretical traditions and reveals a gap in existing pedagogical approaches, which tend to rely on isolated theoretical elements rather than a unified instructional sequence that explicitly incorporates AI as a pedagogical variable. Third, the STAC framework is theoretically modelled as a four-stage instructional sequence. During the Study phase, learners are deliberately exposed to high-complexity materials, with AI as an external regulatory scaffold supporting clarification and cognitive stabilisation outside the classroom. In the Think phase, instructors provide structured conceptual scaffolding within the learners’ Zone of Proximal Development, enabling the systematic closure of comprehension gaps intentionally opened during preparation. The Act phase focuses on experiential application through problem-solving and decision-making tasks, while the Connect phase consolidates learning by explicitly linking theory, action, and reflection to support knowledge transfer to novel contexts.

The paper concludes that STAC provides a theoretically grounded model for integrating AI into the pedagogical process in higher education. Rather than positioning AI as a problem-solving substitute for learners, the framework deliberately preserves the student as the primary cognitive agent and uses AI as a regulatory and support mechanism within the learning process. In contrast to prevailing approaches that incorporate AI mainly as a generic problem solver, STAC demonstrates that meaningful integration requires a structured methodology that raises the level of cognitive demand for both students and instructors. By combining intentional exposure to complex content, AI-supported regulation, and instructor-led conceptual consolidation, the framework argues that a qualitative leap in higher education can only be achieved by increasing—rather than lowering—the intellectual demands placed on all actors involved.

Keywords:

STAC, framework, methodology, AI, Artificial Intelligence.