InjazAI

OpenAI has begun rolling out an invisible, machine-readable text watermark in ChatGPT and Codex outputs, starting with users in the European Union, to meet EU AI Act requirements.

1 min read

OpenAI Adds Text Watermarking to ChatGPT and Codex Under EU AI Act

What happened

FAQ

What is textGrain watermarking in ChatGPT?

It is an invisible, machine-readable watermark embedded in text generated by ChatGPT and Codex, designed to help identify AI-generated content. It is rolling out first to users in the European Union.

How does textGrain compare to Google DeepMind's SynthID?

OpenAI says textGrain \"matched or exceeded\" other approaches like SynthID for text, which is also the basis for the watermarking Anthropic announced in August.

Why is OpenAI starting with the EU only?

The rollout is driven by the requirements of the EU AI Act, the same regulatory pressure that led Anthropic to make a similar move.

Should MENA teams prepare for text watermarking?

Yes. Regional teams should track this development, since any expansion beyond the EU could bring similar disclosure expectations and affect AI-content detection tools and internal publishing policies.

Source: The Verge AI

AI-assisted summary of the source above. High-confidence stories publish automatically; others wait for review. Facts that are not in the source are not added.