OpenAI will watermark ChatGPT text in the EU but makes it optional for API users worldwide
Source: The Decoder (opens in a new tab) · Matthias Bastian
Intel Summary
OpenAI is implementing invisible textGrain watermarks for ChatGPT text in the European Union while allowing API customers worldwide to opt out. Reported testing indicates detection rates reach up to 95 percent, but fall to 17 percent when a quarter of the generated words are substituted.
Why It Matters
The optional watermarking policy gives enterprise API customers flexibility over their output attribution, contrasting with stricter competitor approaches like Anthropic's. However, the steep drop in detection when text is modified underscores technical limitations in relying on watermarking for synthetic text governance and authenticity verification.
Part of an ongoing development
Independent reportingOpenAI implements watermarking for ChatGPT text in the EU
OpenAI is implementing invisible textGrain watermarks for ChatGPT text in the European Union while allowing API customers worldwide to opt out. Reported testing indicates detection rates reach up to 95 percent, but fall to 17 percent when a quarter of the generated words are substituted. Claims are as reported; this summary makes no determination about accuracy or significance.
- Confidence
- Moderate confidence
- Corroboration
- Corroborated
More coverage of this development
- OpenAI will start watermarking ChatGPT’s text in the EUTechCrunchIndependent reporting
Organizations & Entities
Related Intelligence
- ReportSame development
OpenAI will start watermarking ChatGPT’s text in the EU
OpenAI plans to implement invisible text watermarking for ChatGPT and Codex outputs in the European Union to comply with the EU AI Act. The company notes that subsequent text editing can reduce the detectability of the watermarks.
TechCrunch - DevelopmentStableAlso involving Anthropic
Google rolls out Gemini 4 Argon
Google has rolled out Gemini 4 Argon, described as Alphabet's most advanced AI model to date. Claims are as reported; this summary makes no determination about accuracy or significance.
6 independent sources - DevelopmentStableAlso involving Anthropic
Anthropic files S-1 IPO prospectus
According to the Financial Times, Anthropic has submitted its S-1 IPO prospectus, reporting an $8 billion loss on $4.6 billion in revenue over the prior year. The Claude developer's public filing explicitly includes formal risk disclosures warning investors of existential risks to humanity. Claims are as reported; this summary makes no determination about accuracy or significance.
6 independent sources - DevelopmentStableAlso involving Anthropic
Meta to unveil new AI products as Muse tops US app charts
The Financial Times reports that Meta will unveil new AI products as its personal AI agent app, Muse, becomes the most downloaded application in the United States. Claims are as reported; this summary makes no determination about accuracy or significance.
4 independent sources