OpenAI 将在欧盟为部分 ChatGPT 和 Codex 文本输出添加隐形水印

OpenAI says the watermark will roll out to eligible ChatGPT and Codex users on all plans in the EU over the coming weeks. API developers worldwide can enable it for select models starting today, but it is off by default. The watermark survives copying and pasting, though editing weakens detection: in one test, replacing 10% of words with synonyms reduced detection from about 92% to 66%. Initial detector access is limited to approved researchers and expert organizations.

OpenAI says it will begin watermarking eligible text generated by ChatGPT and Codex in the EU over the coming weeks. API customers worldwide can enable it for select models today. The signal lies in word choices and weakens with editing; OpenAI’s detection tool is not available to the public.

OpenAI says it will introduce text watermarks for eligible ChatGPT and Codex users in the EU over the coming weeks, without making them the global default at launch. The watermark embeds an invisible statistical signal during word selection; API customers worldwide can opt in for text outputs from some models starting today.