Researcher Rebuts "Pangram Never False-Positives" Claim With ChatGPT Translation Test
mgaldino · x · 2026-10-23
Responding to Brian Leiter's claim that Pangram never produces false positives, mgaldino reports that an abstract he wrote in Portuguese and had ChatGPT merely translate into English was flagged as 100% AI — offering to share the example. Corroborated by his follow-up post with detection screenshots, it's a direct counterexample to AI detector reliability.
Related event: Pangram AI detector sparks dispute over false positives(2 posts)→
More from Models
- yacine claims OpenAI models have allowed 'borderline and dangerous' tasks since codex 5.4 — yacineMTB · 2026-10-26
- Block unveils Dessa, a full-duplex speech model trained only on human conversation — daluoseo · 2026-10-26
- DeepSeek V4.1F tops Chinese models on FrontierMath Tier 4 at 46% — teortaxesTex · 2026-10-26
- Casado slams Claude's safety classifier for flagging 'fetching my own documents' as reasoning extraction — AccBalanced · 2026-10-26
- Yacine MTB: DeepSeek is 'undoubtedly' a frontier model — yacineMTB · 2026-10-26
- Single-neuron steering: injecting a "joyful neuron" visibly alters a model's self-description — Sauers_ · 2026-10-26