The question is moving from what AI can do to how anyone can check it. Invisible watermarks for eligible ChatGPT and Codex text in the EU over the coming weeks; API opt-in worldwide, off by default. OpenAI's tests show light editing cuts detection from about 92% to 66%.
What OpenAI said it will add
On Monday 5 October, OpenAI said it will add an invisible watermark to eligible ChatGPT and Codex text output in the European Union over the coming weeks, across all plans, to comply with EU AI Act transparency rules that took effect on 2 August. API customers anywhere can opt in for select models from 5 October; it is off by default, and OpenAI said it is not making text watermarking a global default at launch.
The watermark works by subtly shaping word choice, travels with copied text and does not identify the user. OpenAI's own tests suggest editing weakens it: replacing 10% of words with synonyms dropped detection from about 92% to 66%, and short passages, maths answers and translations are harder to detect.
What a watermark does not show
OpenAI also said a watermark "does not measure human contribution," "does not establish ownership or responsibility" and "does not verify accuracy." Anthropic announced global, non-optional watermarking for Claude in August. Access to OpenAI's detector is limited to approved researchers for now. The UK is outside the EU, so UK users are not covered by the ChatGPT rollout; UK firms using the API choose whether to switch it on.
Coverage context
An invisible EU text watermark is now on the record for ChatGPT and Codex, with OpenAI's own tests showing light editing cuts detection, and with an explicit limit on what the mark can prove.
Read the full daily brief hub: www.aifusionautomations.com/news/