# OpenAI begins watermarking ChatGPT and Codex text in the European Union to meet AI Act transparency rules

The textGrain signal reaches about 95% detection at 400 tokens under ideal conditions but falls to 17% when a quarter of the words are swapped for synonyms. API customers anywhere can opt in from October 5.

- Published: 2026-10-06T06:03:25.241Z
- Canonical: https://polylog.news/ai/2026-10-06/openai-begins-watermarking-chatgpt-and-codex-text-in-the-eur
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [OpenAI News](https://openai.com/index/eu-text-provenance), [TechCrunch](https://techcrunch.com/2026/10/05/openai-will-start-watermarking-chatgpts-text-in-the-eu/), [Xenospectrum](https://xenospectrum.com/en/openai-eu-textgrain-watermark-detection-limits/)

OpenAI said it will add an invisible watermark to text from ChatGPT and Codex for users in the European Union (EU). The system, called textGrain, embeds a statistical signal in word choice. Where several words fit equally well, a secret key…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-10-06/openai-begins-watermarking-chatgpt-and-codex-text-in-the-eur (subscription information: https://polylog.news/pricing).