Over 2/3 of paper reviews in the past year were obviously AI-generated, says NLP researcher
bclavie · x · 2026-09-20
NLP practitioner bclavie claims that over the past year more than two-thirds of paper reviews at some institutions were obviously AI-generated with little human input. Because institutions mandate quotas of Corr A/A papers, authors now pre-emptively placate AI-written review complaints, which materially boosts acceptance odds — a game he says they're forced to play. He adds that post-iteration papers are often worse than the original drafts. Quoted skeptic @idansc argues iterating with GPT still can't guarantee novelty or gains over baselines, which reviewers check first.
Related event: Insider Says Over Two-Thirds of Peer Reviews at One Lab Are AI-Generated(2 posts)→
More from AGI Musings
- After AI's rise in math, solvers are told 'not the way we approved' — kjgeras · 2026-09-20
- Jevons paradox is classic low-end disruption you won't spot from a GPU-rich hyperlab — cramforce · 2026-09-20
- Doomer critique: AI apocalypse is an infinitely renewable narrative that never comes due — Dr_Singularity · 2026-09-20
- Schmidhuber: Gauss and Legendre invented 2-layer neural nets back in 1795-1805 — SchmidhuberAI · 2026-09-20
- AI meets math: defensive mathematicians and mean technical people — kjgeras · 2026-09-20
- Capabilities researchers are beyond shame; safety researchers are my audience — RichardMCNgo · 2026-09-20