Policy/press release/2026-10-05

OpenAI implements text watermarking for ChatGPT and Codex in the EU to comply with AI Act

OpenAI has announced its strategy to implement text watermarking for content generated by its AI models, specifically ChatGPT and Codex, for users within the European Union. This initiative is a direct response to the upcoming EU AI Act's requirements concerning content provenance and transparency. The company stated this approach aims to help users identify AI-generated text, although the watermarks are described as invisible. The central claim regarding the implementation of these watermarks originates directly from OpenAI itself, as reported across numerous general and trade publications. While the rollout is initially focused on the EU, some reports indicate that watermarking will be optional for API users globally. The effectiveness and ease of removal of these digital watermarks have been noted by some outlets as potential challenges.

18 articles from 18 outlets covered this story. Their coverage differs on 3 points. The underlying claim is sourced from a press release.

What do all outlets agree on?

18 outlets covered “OpenAI implements text watermarking for ChatGPT and Codex in the EU to comply with AI Act”. All of them report the following:

OpenAI is implementing text watermarking
The watermarks apply to text generated by ChatGPT and Codex
The initial rollout is for users in the European Union
This action is taken to comply with the EU AI Act
The watermarks are described as "invisible" or "digital"

Did outlets disagree about this?

Yes. Coverage of “OpenAI implements text watermarking for ChatGPT and Codex in the EU to comply with AI Act” differs on 3 points. Each account below is how a different outlet described the same event:

The watermarks are not hard to remove

Neowin

Watermarking is optional for API users worldwide

the-decoder.com, The New Stack

OpenAI's approach to watermarking differs from Anthropic's

The New Stack