Generative AI in Art and Design Education: Keeping the Human in the Creative Loop

Research on generative artificial intelligence in education reveals a persistent divergence between successful assisted performance and the development of independent knowledge, creative judgment, and transferable skill. Studies of art and design additionally identify opportunities for iterative exploration alongside concerns about visual fixation, uncertain authorship, and dependence on generated alternatives. These findings create an instructional problem: educators need to determine which creative decisions should remain with learners, which operations can receive bounded assistance, and which activities can be delegated without bypassing the intended learning. Existing scholarship addresses these questions through research on scaffolding, human-AI collaboration, assessment, and psychological ownership, but these contributions require closer integration within everyday studio practice. This article develops a Process-Centered AI Integration Framework for secondary and postsecondary art and design courses. The framework allocates tasks through three modes, protect, support, and delegate, according to learning objectives and demonstrated evaluative capability rather than the proportion of an artifact generated by AI. Six recursive phases connect intention formation, bounded assistance, comparison of alternatives, purposeful revision, human critique, and reflection with independent transfer. Parallel curricular applications, assessment criteria, and a reusable assignment planner specify retained decisions, appropriate AI roles, and concise evidence of learning. The model addresses the performance-learning distinction by making disciplinary reasoning observable across creative production while treating process ownership, product ownership, and achievement as related but distinct educational concerns.