OpenAI is implementing invisible text watermarks in ChatGPT and Codex, initially for users within the European Union, to comply with the EU's Artificial Intelligence Act. While the EU rollout is mandatory, OpenAI is offering an opt-in option for its API customers globally. The company claims its textGrain watermarking technology performs comparably to similar methods from Google DeepMind and Anthropic, though it acknowledges detection is not guaranteed and does not verify accuracy or authorship. AI
IMPACT This feature aims to increase transparency and aid in identifying AI-generated content, potentially impacting content moderation and trust in AI outputs.
RANK_REASON This is a product feature rollout for an existing AI model, not a new model release or significant policy change.
Read on Mastodon — sigmoid.social →
- Anthropic
- application programming interface
- Artificial Intelligence Act
- ChatGPT
- codex
- EU
- European Union
- Google DeepMind
- OpenAI
- SynthID
- textGrain
- The Verge
AI-generated summary · Google Gemini · from 6 sources. How we write summaries →