Effects of LLM Use and Note-Taking on Reading Comprehension and Memory: A Randomised Experiment in Secondary Schools

A randomised experiment in secondary schools by a Microsoft Research and Cambridge team: 405 pupils aged 14–15 read a text under one condition — with an LLM, taking notes, taking notes plus an LLM, or unaided. Comprehension and retention were then tested.

Note-taking — alone or combined with the LLM — produced significantly better comprehension and memory than the LLM alone. The LLM alone came last. Yet pupils clearly preferred the LLM and rated it more helpful than note-taking.

That gap between feeling and result is the point: a tool that feels like help measurably weakens what it was meant to serve — and the pupil does not notice. For parents and teachers it means that what a child finds helpful cannot be trusted as a guide. The practical lesson: LLM plus one’s own work works; LLM instead of it does not.