OpenAI to Implement Invisible Watermark for ChatGPT Text in the EU

Oct 5, 2026

AI Summary

OpenAI will introduce an invisible watermark for text generated by ChatGPT and Codex in the European Union to comply with the EU AI Act's transparency requirements. This watermark will help identify AI-generated content without altering its performance and will be available to developers using OpenAI's API globally, though it will be off by default.

OpenAI announced it will add an invisible watermark to text generated by ChatGPT and Codex in the EU to comply with the EU AI Act.
The EU AI Act's transparency rules require AI-generated content to be identifiable by other systems.
The watermark will be rolled out to eligible users in the EU over the coming weeks, while developers globally can enable it for select models starting now.
The watermark subtly alters word choices to create an undetectable pattern that can be identified by detectors, and it remains with the text when copied.
OpenAI stated that the watermark does not identify users and does not significantly affect model performance.
A technical report detailing the watermarking method, called textGrain, was published in collaboration with researchers from the University of Pennsylvania and Yale.
Tests indicated that editing the text could reduce detection effectiveness, with significant changes dropping detection rates considerably.
OpenAI emphasized that a missing watermark does not confirm human authorship, as text could be too short or heavily edited.
The decision to watermark follows a similar move by Anthropic, which faced user backlash for its approach to watermarking AI-generated content.
Other major AI companies, including Google, Meta, and Microsoft, have also committed to adhering to the EU's code of practice on AI-generated content.

openaiwatermarkingai actchatgpteu

OpenAI to Implement Invisible Watermark for ChatGPT Text in the EU

Related Stories

Former Biosecurity Official Highlights AI Risks and Calls for Stronger Safeguards

UK Government to Limit Non-Compete Clauses Following Startup Concerns

AI Companies Hire Former Trump Officials to Strengthen Ties with U.S. Government

Japan Calls for Security Reviews Amid Surge in Cyberattacks on Companies