cognitive debt
In plain wordsThe idea that letting AI do your thinking or writing saves effort now but weakens your own thinking, memory and learning over time.
The hypothesized long-term cost to a person's thinking, memory and learning that builds up when they habitually hand cognitive work to AI assistants — by analogy with technical debt, effort saved now is repaid later as weaker skills and understanding. The term comes from an MIT Media Lab study in which 54 participants wrote essays across several sessions, either unaided, with a search engine or with a chatbot, while EEG recorded their brain activity. The chatbot group was least able to quote its own essays, felt the least ownership of them and showed the weakest neural connectivity, and some differences persisted when its members later wrote without help. Critics point to the small sample, the narrow task and the paper's preprint status, and warn against reading it as proof of lasting harm.
The teacher banned chatbots for first drafts, citing cognitive debt rather than cheating.
Origin
Introduced by Nataliya Kosmyna, Pattie Maes and colleagues at the MIT Media Lab in the preprint 'Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task' (arXiv, 10 June 2025), which drew wide media attention.
Usage note
Related to the older psychological idea of cognitive offloading — relying on external aids such as notes or devices to reduce mental effort — which is not harmful in itself; see deskillingdeskilling, n.: The loss of people's skills when machines take over the work that used to keep those skills sharp, or when jobs are redesigned so that less skill is needed..
Related
- deskillingdeskilling, n.: The loss of people's skills when machines take over the work that used to keep those skills sharp, or when jobs are redesigned so that less skill is needed.
- overrelianceoverreliance, n.: Leaning on an AI system more than its reliability justifies: accepting its answers without checking, or letting it make decisions it often gets wrong.
- automation biasautomation bias, n.: The habit of trusting a computer's suggestion too much: following it even when it is wrong, or missing problems because the machine did not flag them.
- AI literacyAI literacy, n.: Knowing enough about AI to use it sensibly and judge it critically: what it can and cannot do, how it goes wrong, and how it affects people.
- workslopworkslop, n.: AI-generated work that looks polished but does not actually get the job done, quietly passing the real effort to whoever receives it, who must then puzzle it out or start over.
- augmentationaugmentation, n.: Using AI to help people do their work better or faster while they stay in charge, instead of having the AI do the job on its own.
Sources
- Kosmyna, N., … Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. arXiv:2506.08872. arxiv.org/abs/2506.08872