Creative offloading and creative agency in AI-assisted educational game design: A thematic analysis of student reflection essays

Generative AI has raised growing educational concerns regarding cognitive offloading and the preservation of students’ creative agency in creative learning tasks. This study examined how higher education students reflected on their creative agency when using generative AI as a design partner during the ideation and initiation phase of an educational game design assignment. Using a qualitative case study approach, we conducted thematic analysis of reflection essays written by 26 students following an AI-supported game design activity in a university course. The analysis identified a spectrum of reflected creative agency patterns related to how students described delegating, evaluating, and directing creative work with generative AI. Students’ reflections described strategic forms of offloading in which AI was positioned as a complementary support tool while evaluative judgment and authorial direction were retained. At the same time, reflections also revealed more unregulated forms of offloading characterised by overreliance on AI, unclear interaction strategies, and diminished ownership of creative decisions. These patterns sometimes coexisted within individual reflections, suggesting that creative agency may vary across offloading decisions rather than represent fixed student characteristics. The study introduces the concept of creative offloading as a lens for understanding student-AI interaction in creative educational tasks.