From the source
OpenAI is rolling out text watermarking for API customers globally and ChatGPT and Codex users in the EU to comply with the EU AI Act's requirement that generated text be machine-identifiable.
- The company says text watermarking and detection remain early technologies with significant limitations, and its phased approach reflects both the legal requirements and the technology's current constraints.
- Starting today, API customers can opt in to watermarking for select models, while ChatGPT and Codex output in the EU will receive invisible watermarks over the coming weeks.
- OpenAI is also opening applications for its text watermark detector to approved researchers and expert organizations, citing evaluations showing detection drops sharply with shorter text, constrained word choice, or editing.





