Stanford prof backs proposal for official LLM reviews in peer review, humans just calibrate

anshulkundaje · x · 2026-09-29

Stanford professor Anshul Kundaje and suragnair discuss using LLMs to reform academic peer review. suragnair proposes creating an "official LLM/agent review" for each paper, so reviewers without much substantive feedback can simply comment on and calibrate the AI's review.

Kundaje fully agrees. The idea targets lazy reviewing head-on: instead of reviewers quietly offloading to LLMs, make the AI review public and have humans supervise it.

Original post →

More from Research

Research channel →