ToolsThe story, in brief
OpenAI will start watermarking ChatGPT’s text in the EU
OpenAI is watermarking ChatGPT outputs in the EU — but the marks fade if you edit. Here's what that means for compliance.
The KeyNews take
Why it matters
OpenAI is deploying invisible watermarks on ChatGPT and Codex text to satisfy EU AI Act provenance rules. The marks degrade under editing, creating a compliance gap that EU regulators will likely scrutinize.
The key facts
5 to knowWatermarking applies to ChatGPT and Codex text in EU only
Marks are invisible to end users
Editing reduces detectability of watermarks
Implementation driven by EU AI Act compliance requirements
No timeline for rollout or technical specification disclosed
The story so far
Earlier coverage of this storyline
Go to the source
TechCrunch AItechcrunch.com
Publisher excerpt: OpenAI will watermark ChatGPT and Codex text in the EU to comply with the AI Act. Editing can make the invisible marks harder to detect, it says.