AI-Aesthetic TPACK: Redesigning Art Teacher Education in the Age of Generative AI

Abstrak

Generative text-to-image systems lower technical barriers to visual production but can also short-circuit learning if used end-to-end. This paper proposes AI-Aesthetic TPACK, a framework that aligns Human-in-the-Loop (HITL) and Human-out-of-the-Loop (HOTL) workflows to Leder’s five stages of aesthetic processing (perceptual analysis, implicit memory integration, explicit classification, cognitive mastering, evaluation) and operationalizes it as a classroom toolkit: (a) a behavior-anchored competency rubric for teacher education; (b) a Process-Based Evidence Package (PBEP) template that makes creative processes visible and assessable; and (c) a course-level AI use & disclosure policy. Using a PRISMA-ScR–guided scoping synthesis, we distill design patterns that preserve stage-specific learning while leveraging AI for ideation and iteration. A 90-minute micro- trial with three experienced art instructors shows moderate inter-rater agreement when applying the rubric to a PBEP-documented student project (κoverall = 0.65; κdimension = 0.58–0.73). Qualitative debriefs reveal consistent pain points (e.g., ambiguity around licensing of texture assets) and lead to a shared casebook for future calibration. Contributions are: (1) a stage-aligned mapping of AI interventions with scaffold/shortcut conditions; (2) a ready-to-use toolkit (rubric + PBEP + policy) with evidence anchors; and (3) initial usability/consistency signals for short-format teaching. We conclude with boundary conditions, cultural considerations, and a data/materials availability statement to support replication and local adaptation.