OpenAI plans EU text watermarking for ChatGPT and Codex, opens global API opt-in

OpenAI reports no meaningful performance difference with and without watermarking across its Astra benchmarks. It plans to release the technology as open source.

Detection has limits

OpenAI is opening detector-access applications, initially for approved researchers and expert organizations rather than the public. Its evaluations show why interpretation needs care: in a test of 400-token watermarked English responses to questions from the ELI5 dataset, replacing 10% of words with synonyms reduced detection from about 92% to 66%; replacing 25% reduced it to 17%.

The company also says short or more constrained text is harder to detect. A watermark does not establish ownership, accuracy or the amount of human contribution. Conversely, failing to detect one does not prove a passage was written by a person.

Replies largely criticized OpenAI's textGrain watermarking rollout for ChatGPT in the EU, calling it intrusive and ineffective, while some welcomed it for helping expose disguised AI-generated content.

Based on 636 sentiment-bearing replies from 509 accounts across 13 conversations.

Replies largely criticized OpenAI's textGrain watermarking rollout for ChatGPT in the EU, calling it intrusive and ineffective, while some welcomed it for helping expose disguised AI-generated content.

Based on 636 sentiment-bearing replies from 509 accounts across 13 conversations.