MIT: Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task

“These results raise concerns about the long-term educational implications of LLM reliance and underscore the need for deeper inquiry into AI's role in learning.…”

This study examines how the use of large language models (LLMs) like ChatGPT affects neural, cognitive, and writing behaviour over repeated tasks. Participants were divided into three groups: using an LLM, using a search engine, or using no tools (brain only). They performed essay writing across multiple sessions while their brain activity was monitored via EEG. The findings show that LLM users exhibited weaker neural connectivity, reduced cognitive engagement, and lower self-reported ownership of their writing. When those same users were later switched to tool-free conditions, their performance and brain activation patterns declined relative to peers.

The results raise significant concerns about the long-term educational effects of overreliance on AI writing assistants. While LLMs provide convenience and short-term benefits, the study suggests they can create what the authors call “cognitive debt,” weakening learners’ ability to think, write, and engage independently. The authors advocate for a more intentional, scaffolded approach to AI integration in education to ensure that human cognitive development remains central to learning.