Issue 2026-11-17 · Industry · 研究
Researchers report AI pre-submission reviews return 'borderline' verdicts
A Reddit user shared that when they use Claude or ChatGPT to pre-review CVPR submissions, the AI always returns a 'borderline' verdict with 'limited novelty' no matter how many fixes are made, sparking community discussion.
Read original ↗- Major Data Breaches at Gemini, Claude, and ChatGPT Could Put Millions at Risk
- Developer launches Fakemetrics to fake Stripe growth screenshots
- Politico tests energy news inside Claude, ChatGPT and Copilot
- New French Prix Pascal prize to reward AI-assisted writing
- Can AI chatbots replace mental health therapy? Experts say no