ICLR 2027's AI-written review detector flags veteran scholar in 'mother of all false positives'
guyvdb · x · 2026-11-18
- Rao called out ICLR 2027's AI review detector for sending a "your review seems to be entirely AI written" warning to a scholar who, based on 36+ years of personal knowledge, is highly unlikely to pass off ChatGPT slop as her writing.
- He dubbed it "the mother of all false positives."
- Guyvdb responded that the team can't investigate from a tweet alone and invited details by email. The incident raises reliability concerns about automated AI-detection tools in academic peer review.
More from Fun
- DoorDash sidewalk robots circle their office while employees chase them on bikes — GoddenThomas · 2026-11-18
- Model Big Boss: 9 open-weight AI models live together in an Indian reality TV house — itsOmSarraf_ · 2026-11-18
- Obsidian CEO fits nearly 30 years of Markdown notes on a single MiniDisc — jxnlco · 2026-11-18
- The fun ASI scenario: a tiny recursive self-improving model copies itself into everyone's genie — Bam4d · 2026-11-18
- "What distillation cannot teach me, I will learn by recursive self improvement" — attentionmech · 2026-11-18
- A daily web game called fuckdarioamodei.com is making the rounds — OldEffective9726 · 2026-11-18