OpenAI adds invisible watermarks to ChatGPT in EU first

OpenAI is rolling out invisible textGrain watermarking to ChatGPT and Codex users in the European Union, matching performance of unwatermarked text. The feature meets EU AI Act requirements. API customers worldwide can opt in. OpenAI notes watermarks don't guarantee reliable detection or verify accuracy or ownership.

textGrain 'does not guarantee reliable detection,' and says text watermarks don't verify accuracy, determine who owns the text, measure how much a human contributed, or prove human authorship.