Artificial CreativityConsumer Behavior
I am super happy to this this out: A chapter I co-wrote with my dear colleague Jessie de Witt Huberts, published on 30 October 2026 in The Oxford Handbook of Human Creativity and Generative Artificial Intelligence in Education, edited by Ronald Beghetto. In our chapter, titled “Cognitive Offloading to Artificial Intelligence in Education: A Critical Analysis of the Long-Term Effects“, we ask a question most AI-in-education discussions skip: what happens to people’s thinking when AI does the thinking for them, not for one assignment, but over years?
What is cognitive offloading to AI?
Cognitive offloading is handing mental work, such as recalling, structuring or evaluating, to an external tool so you no longer do that work yourself. Calculators are the classic case. Generative AI applies the same mechanism to almost any intellectual task, which is why the stakes are different now.
Think of a GPS. Using it to cross town is fine. The problem starts when you never learn the streets, because then you cannot tell when the app sends you the wrong way. The tool is not the issue. The skill it replaced was never built.
What the chapter argues
- The premature and uncritical adoption of AI in formal education risks undermining the core cognitive processes that long-term learning depends on.
- We propose a conceptual model in which human agency and AI-driven offloading move in opposite directions, and that trade-off shapes ownership and cognitive engagement.
- Grounded in self-determination theory and psychological ownership, the chapter looks at how AI shifts creative agency, and how that can affect intrinsic motivation, self-efficacy and creative identity.
- It closes with policy recommendations and a future research agenda.
Why the inverse relationship matters
In plain terms: the more a student lets AI carry the work, the less that student owns the result and the less engaged they are with it. Ownership is the felt sense that something is yours. It is part of why effort matters, not just the output.
A student who did not wrestle with a problem cannot defend its answer under questioning. That answer never becomes part of how they think. Multiply that across a degree programme and you get graduates who can produce polished work they cannot fully explain.
Short-term memory will not solve complex problems
Complex problems require a stock of knowledge you can reason with without a tool in your hand. An answer you can retrieve is not the same thing as a foundation you can build on.
A chef who has only ever followed recipes from a tablet will struggle the day the oven breaks. The recipe was never the skill. Judgement was, and judgement comes from doing the work often enough to get it wrong.
Balance, not abstinence
Our position is not that schools should ban AI. Most learning should still happen without it, because foundations have to exist before augmentation makes sense. Once they do, AI can extend what a student can do. The real question is sequencing and purpose: when does a student work unaided, and when do they use AI, and for what?
In practice, that looks like:
- Attempting the problem before opening the tool.
- Requiring students to explain and defend any AI-assisted output in their own words.
- Assessing the reasoning process, not only the final product.
Accountability is the missing piece
Without accountability, AI use becomes invisible substitution. Two kinds matter. Formal accountability means assessments that check the student’s own reasoning, such as in-class work, oral defences and process documentation. Informal accountability means teachers and peers asking students how they got to an answer, every day, not only at exam time.
Who gets ahead with AI
The people making the most of AI in education, and outside it, are the ones with a strong base in their field. A calculator helps a mathematician. It does not teach arithmetic to someone who never learned it. AI works the same way: it amplifies expertise, and it cannot manufacture expertise that was never built.
That is the argument for keeping foundations central. It is not nostalgia for pre-AI classrooms.
Read the chapter
The full chapter is in The Oxford Handbook of Human Creativity and Generative Artificial Intelligence in Education (Oxford Handbooks, online 30 October 2026; print January 2027). It is behind institutional access, so check with your library if you cannot reach it.
- Chapter DOI: https://doi.org/10.1093/9780197832431.003.0007
- Chapter page: https://academic.oup.com/edited-volume/63609/chapter-abstract/583705121
If you work in education, the question to take from this is not whether students use AI. They will. The question is whether anyone is checking that they can still think without it.
Protect your cognitive sovereignty.
