Home · Sources · Contractor & Reyes 2026
Experimental Evidence on the Learning Impact of Generative AI
Contractor & Reyes 2026 · randomised trial · touches 1 claim · used in 1 attack · used in 1 correction to our notes
Read the study itself (doi.org) ↗
Which beliefs it touches
- C2 Answer-giving AI in practice sets a limit on it
Off-the-shelf GenAI access: delayed unaided knowledge tests +0.21 SD at ~1 week in v2 (p = .068; v1 said +0.27 SD), no loss; scopes C2 to outsourcing / unprotected answer-giving (weakened 2026-10-03).
- C2 Answer-giving AI in practice pushes against it
Overall ITT positive for typical undergraduate GenAI access — kill of typical-GenAI-lowers-retention prong.
Attacks that used it
- C2 attack #5, 3 October 2026: narrowed
Corrections to our notes that used it
- C2, 16 October 2026: corrected