OPENAI TO WATERMARK CHATGPT TEXT - HERE WHY

What happened: OpenAI said it will add an invisible watermark to text generated by ChatGPT and Codex for European Union users to comply with the EU AI Act, rolling out over the coming weeks to eligible users on all plans in the EU. Key numbers: Replacing about 10% of words with synonyms cut detection accuracy from about 92% to 66% in one company test. Why it matters: OpenAI is not making watermarking a global default at launch, and detector access goes only to approved researchers and expert organisations. Anthropic announced in August it would watermark Claude text globally.

OpenAI watermarking ChatGPT text signals AI content provenance amid regulatory scrutiny.

OpenAI watermarking ChatGPT text signals AI content provenance amid regulatory scrutiny. — impact context not separately stored; see What changed.