OpenAI launches invisible text watermarking for EU, unveils textGrain technique
机器之心 · wechat · 2026-10-06
OpenAI enabled text watermarking: API customers can opt in globally, and EU-bound ChatGPT/Codex outputs get invisible marks to comply with the EU AI Act Art. 50. The textGrain report (senior author: COPSS award winner Weijie Su) introduces an entropy budget and block-wise optimal transport to balance detectability against output diversity. Benchmarks show no significant quality drop, but detection falls to 17% after 25% synonym substitution. Detector access is limited to vetted researchers; report to be open-sourced.
Related event: OpenAI launches textGrain text watermarking, starting in the EU(25 posts)→
More from Models
- Google silently disables logprobs on Gemini Hydrogen, Neon, Argon and Carbon models — dejanseo · 2026-11-20
- Bug Hunt chart: score vs subscription cost, Gemini 4 Neon xhigh tops at 63 for $1.80 — PawelHuryn · 2026-11-20
- Measured: Claude Max delivers 46.4x API value on Opus 5.5, beating ChatGPT Pro's 10.25x — PawelHuryn · 2026-11-20
- Dev complains GPT keeps skipping steps and ignoring instructions — lxfater · 2026-11-20
- GPT-6.2 Astra and Sol show OpenAI optimizes performance per dollar, not raw capability: analysis — VraserX · 2026-11-20
- Mamba+MoE hybrid reasoning model Olala-v1.0-mini-7A1B trends on Hugging Face — OVHaiLLM · 2026-11-20