ICML 2027 chair crowdsources fixes for AI slop; researcher proposes LLM desk-reject of 90% papers

ZeeshanZiaML · x · 2026-09-21

ICML 2027 Program Chair Andrew Wilson asked for creative ideas on handling exploding submissions, AI slop, reviewing quality and bureaucracy. Zeeshan Zia proposed a concrete workflow: build a 50-accept/50-reject test set at the PC level, tune an LLM prompt against it, then use that prompt for a first desk-rejection pass targeting 90% desk rejection.

Related event: ICML 2027 Chair Seeks Ideas as AI Slop and Submission Surge Threaten Peer Review(8 posts)→

Original post →

More from Companies & People

Companies & People channel →