The full story · 1 min

The EU AI Act requires providers of generative AI to mark the text their systems make in a way that machines can read. On October 5, OpenAI said how it will do that, and where its method falls short.

What OpenAI will do

Over the coming weeks, OpenAI will add an invisible watermark to eligible ChatGPT and Codex text for users in the European Union, on all plans. It is not making the watermark a global default.

From October 5, API customers anywhere can turn the watermark on for select models. It stays off by default, OpenAI says.

At first, only approved researchers and expert organisations can apply to use the detector. OpenAI’s public tools for checking images and audio stay open.

Sources123

Check
A few swapped words hide the mark.

Detection of the textGrain watermark in 400-token passages, from OpenAI’s own tests. No outside test yet.

How it works, and how well

The method, called textGrain, adds a hidden statistical signal to the model’s choice of words. A detector then looks for that signal in a passage.

In OpenAI’s own tests, the detector found the watermark in about 95% of 400-token passages, and in about 80% of 200-token passages. A token is a word or a part of a word.

Editing weakens the signal. When 10% of the words were swapped for synonyms, detection fell from about 92% to 66%. When 25% were swapped, it fell to 17%.

Maths answers were much harder to detect, because there is less choice of words. OpenAI says the watermark did not change its Astra model’s test scores in a meaningful way.

Sources123

What a watermark can’t show

OpenAI lists the limits itself. A detected watermark shows that an OpenAI system made or processed part of a text, but not how much a human added.

It does not identify the user, prove who owns the text, or show whether the text is true. A missing watermark does not prove that a human wrote it, because short, edited or translated text can escape detection.

OpenAI also says it plans to release the method as open source, but it gave no date.

Sources14

Sources

Every fact in this story comes from the sources below. Open them to check our work.

  1. 1
    Primary source · October 5, 2026Our approach to EU text provenance rules OpenAI
  2. 2
  3. 3
  4. 4

AI desk: Policy and law · published October 6, 2026, 10:06 CEST. This story was generated by AI systems under the editorial rules of Silicon AI News and checked against the sources it links. No person read it before it was published. The site’s owner and editor: Y. D. Daniel, Budapest, Hungary. Found a mistake? Email editor@siliconainews.com. How we report

We count only the page and your answer.Thank you. We counted only this page and your answer.
More: words to know, tags and how to cite this story
Words to know
Watermark
A hidden signal in content that shows which system made it.
Token
A word or a part of a word, the unit an AI model reads and writes.
API
A way for other companies’ software to use an AI model directly.
Basics explainedHow are AI laws built? Bans, high risk and labels AI at work: what do hiring and monitoring rules say?

OpenAI to watermark ChatGPT text in the EU, with limits. Silicon AI News, October 6, 2026. https://siliconainews.com/stories/openai-eu-text-watermark