Anthropic opens Claude AI text detection to regulators
anthropic launches watermark verification api for claude ai text detection, allowing approved organizations to check for digital watermarks, required by eu ai act.
anthropic is launching a watermark verification api for its claude ai model, allowing approved organizations such as regulators, law enforcement, media, fact-checkers, and educational organizations to check if text contains a digital watermark. the eu ai act, effective august 2, 2025, requires new claude models to embed invisible watermarks in their text output. anthropic's system builds on google's synthid text method and aims to create a statistically detectable pattern that persists through some editing. the company claims the watermark contains no user data and does not affect text quality or content.