OpenAI announced it will start adding an invisible watermark to text generated by ChatGPT and Codex in the European Union to comply with the EU AI Act. The company said the watermark will roll out over the coming weeks to eligible users in the EU.
The watermark, called textGrain, works by subtly shaping the model’s word choices, leaving a pattern readers can’t see but a detector can pick up. It is not an actual symbol and travels with the text when it’s copied and pasted.
"These limitations contribute to our decision to provide initial detector access only to approved researchers and expert organizations," said the company. OpenAI also cautioned that a missing watermark "does not prove human authorship."
The announcement follows Anthropic’s decision to watermark text generated by Claude, a move it’s applying worldwide. OpenAI had built a text watermark before but held off on releasing it, partly over concerns that users would switch to rivals that didn’t watermark.
OpenAI said the watermark doesn’t identify the user and saw no meaningful change in its models’ performance with it switched on. Can the watermark be removed by editing? OpenAI’s tests suggest yes.
Source: techcrunch
