OpenAI Adds Watermarks to ChatGPT Text by Default in the EU to Comply with the EU AI Act
Understand the story
On October 6, 2026, The Verge and TechCrunch reported that OpenAI launched an invisible text watermark called textGrain for ChatGPT and Codex users in the EU, to comply with the EU AI Act transparency rules effective August 2. The watermark covers all plans and will roll out to EU users over the coming weeks; it will not be enabled globally by default for now. Starting immediately, developers worldwide can opt in to watermarked output via the API for select models, with it off by default. Approved researchers and expert organizations can apply for the detector, which only reports whether an OpenAI watermark is detected, without identifying users or revealing conversation content. On October 7, Ars Technica reported that OpenAI will automatically watermark ChatGPT-generated text by default in the EU, while other regions can opt in but it will be off by default. textGrain embeds watermarks through word-choice patterns imperceptible to the human eye; the detector will first be opened to a limited set of researchers through an application process. Ars Technica also noted that such watermarks remain relatively easy to circumvent for anyone with basic technical skills.
AI overview of reporting · Updated 11 days ago
Reporting timeline
Explore different aspects of the story through its reporting.
- Ars Technica · AIOpenAI will watermark ChatGPT outputs by default—but only in the EU
OpenAI announced it will automatically watermark ChatGPT-generated text by default in the EU, with opt-in elsewhere, to comply with the EU AI Act taking effect in August. The proprietary method, called textGrain, embeds watermarks via word-choice patterns imperceptible to humans; a detector will first be opened to a limited group of researchers through an application process, but the authors note such watermarks remain easy to circumvent for anyone with basic technical skills.
- TechCrunch · AIOpenAI will start watermarking ChatGPT’s text in the EU
OpenAI announced invisible watermarks for text generated by ChatGPT and Codex in the EU to comply with EU AI Act transparency rules effective August 2, rolling out to EU users on all plans in the coming weeks; developers worldwide can enable it in the API for some models starting today, off by default.
- The Verge · AIOpenAI is adding text watermarking in ChatGPT and Codex
OpenAI is rolling out the textGrain invisible text watermark to ChatGPT and Codex users in the EU, covering all plans, though it is not yet enabled globally by default. API customers can opt in to watermarked output for some models starting today, and approved researchers and expert organizations can apply for the detector; the tool only reports whether an OpenAI watermark is detected, without identifying users or exposing conversation content.