The “guided intervention” model: a TPACK-based classroom observation of teacher orchestration in the “Teacher–Student–AI” triad

This qualitative case study investigates how an instructor in a higher education public management course orchestrates generative AI within the “Teacher–Student–AI” triad to foster critical thinking, support students’ ability to independently evaluate AI-generated knowledge, and develop higher-order thinking that transcends AI outputs – collectively referred to as epistemic agency, the capacity to exercise independent evaluative judgment over knowledge claims including those produced by AI. Grounded in an adapted TPACK framework – augmented with the critical-integrative dimensions of scaffolding interaction, critical filtering, and fostering transcendence – the study draws on classroom observations, stimulated recall interviews, and instructional artifacts. Findings reveal three core pedagogical strategies that define the teacher’s role as an orchestrator: scaffolding human–AI co-creation through structured interaction design, critically filtering AI-generated content, and fostering cognitive transcendence beyond AI outputs through a structured learning cycle. These strategies collectively constitute a “Guided Intervention” model, operationalized as the “Explore–Qualify–Create” instructional design heuristic. The study contributes to the discourse on human–AI collaboration by offering an empirically grounded, behaviourally operationalised account of how TPACK-based competencies are enacted in authentic triadic classroom interactions, and by offering actionable guidance for teachers transitioning from technology users to orchestrators of AI-enhanced learning environments.