OpenAI is rolling out an invisible, machine-readable text watermark for outputs from ChatGPT and Codex, starting with users in the European Union. The company says the watermark is designed to help systems identify AI-generated text and is being introduced over the coming weeks.
Outlets describe the move as part of OpenAI’s compliance with the EU AI Act, which requires providers of generative AI to mark AI-generated content in a way that makes it identifiable. 9to5Mac and others emphasize that OpenAI also details how it will apply the approach across ChatGPT, Codex, and its API, tied to the legal requirement.
Several reports note that the watermark is not intended for casual human reading, but for machine detection. TechCrunch adds that editing may make the marks harder to detect. The Verge also places OpenAI’s approach in context by comparing it to similar text watermarking work, including Google DeepMind’s SynthID and Anthropic’s previously announced system, which is said to be based on SynthID.
