ICLR 2027 AC: 10 of 24 reviews flagged AI-generated, 11 reviewers silent
rao2z · x · 2026-11-22
An ICLR 2027 Area Chair reports on their batch of 14 papers: 4 desk-rejected, 2 withdrawn after reviews, and 10 of the 24 remaining reviews flagged as AI-generated (7 serious and unacknowledged). 11 of 24 reviewers never responded to rebuttals, leaving the AC to read all eight papers themselves — a first-hand quantification of AI-generated reviews undermining peer review.
More from Safety
- Right to Compute manifesto pays off: Montana SB212 signed into law — nptacek · 2026-11-22
- NYT: AI Is Exposing Decades of Vulnerable Code as Tech Outages Pile Up — chrismattmann · 2026-11-22
- AI Safety Talk Turns Practical: Call for US-China Team-Up Against Cyber Swarms — shakoistsLog · 2026-11-22
- Microsoft patents AI that rewrites livestream speech and lip movements in near real time — TinfoilTricorn · 2026-11-22
- Researcher BlancheMinerva: OpenAI may be hoarding secret models that could change the world — BlancheMinerva · 2026-11-22
- Zero data retention is trending: ZKPs for model computation go mainstream — yenkel · 2026-11-22