ICML 2027 chair crowdsources fixes for submission explosion; proposal for AI-only review track

thegautamkamath · x · 2026-09-22

ICML 2027 Program Chair Andrew Wilson publicly asked for ideas on handling AI slop, exploding submissions, reviewing load, and bureaucracy. Quoting him, @galyo proposes a radical two-track system — citing 60k ICLR submissions as evidence the community is too slow — where Track B papers are reviewed entirely by AI: conferences still set acceptance standards, but enforcement shifts from human reviewers to an AI review process designed by each conference's steering committee.

Related event: ICML 2027 Program Chair Seeks Community Ideas on AI Slop and Peer Review Crisis(7 posts)→

Original post →

More from AGI Musings

AGI Musings channel →