Paper rejected 2/2/3/2 as AI detector flags 3 of 4 reviews as 100% AI-generated
nsaphra · x · 2026-11-06
Researcher nsaphra reports her potentially most practically impactful paper scored 2/2/3/2, with Pangram flagging 3 of 4 reviews as 100% AI-generated (the shortest review was short and mostly AI-generated). She laments that the community likely can't afford to ban such reviewers from refereeing.
More from Safety
- Subliminal learning happens at every LoRA rank and full fine-tuning, given right learning rate — thebasepoint · 2026-11-07
- Bombay HC flags ChatGPT-drafted petition; advocate withdraws it — kingslyj · 2026-11-07
- Anthropic reportedly tried to poach a METR evaluator mid-project, sparking backlash — peterwildeford · 2026-11-07
- Dutch government runs world's first AI compute verification test on state cluster — Miles_Brundage · 2026-11-07
- "Humans can't write secure infra": Daybreak fields red/blue teams with GPT Cyber models — jxnlco · 2026-11-07
- METR: getting nothing would be worse than accepting Anthropic's eval terms — S_OhEigeartaigh · 2026-11-07