OpenAI unveils a text-provenance solution under the EU AI Act and launches the textGrain text watermark
Our approach to EU text provenance rules
AI BriefingOpenAI unveiled a text-watermarking solution for the EU AI Act and released textGrain, a technology that adds invisible statistical signals to the model's word choices. API customers can selectively enable watermarks for some models starting today. Over the coming weeks, invisible watermarks will be added to ChatGPT and Codex outputs in the EU. The detector is currently available only to approved researchers and expert institutions.
Reported by 9 additional sources
- OpenAI will enable watermarks for ChatGPT outputs by default in the EU and disable them by default elsewhereArs Technica: AI · October 7
- OpenAI will add the invisible textGrain watermark to ChatGPT text in the EUTechCrunch: AI · October 6
- OpenAI launches the textGrain text watermark in ChatGPT and Codex, initially for users in the EUThe Verge: AI · October 6
- OpenAI will add invisible watermarks to ChatGPT and Codex text in the EURohan Paul · October 6
- OpenAI watermarks ChatGPT and Codex text to comply with the EU AI ActOpenAI · October 6
- OpenAI explains how text watermarking works and its capability boundariesOpenAI · October 6
- OpenAI explains the limitations of text watermarking; the detector is currently available only to approved researchersOpenAI · October 6
- OpenAI will add the invisible textGrain watermark to ChatGPT text in the EU; API users worldwide can optionally disable itThe Decoder: AI News · October 6
- OpenAI will add invisible watermarks to ChatGPT and Codex text outputs in the EUIT Home · Artificial Intelligence · October 6
Source: OpenAI: Official Website Updates · openai.com
Event follow-up1
View all event coverage- OpenAI unveils a text-provenance solution under the EU AI Act and launches the textGrain text watermarkOpenAI: Official website updates · October 5