Google opens SynthID Detector worldwide for partner watermarks

Google opened its expanded SynthID Detector to everyone worldwide in English on October 7. The public tool lets users upload images, video and audio to check for supported SynthID watermarks.

The detector can identify watermarks in media made with AI systems from Google and participating partners. Google lists OpenAI, NVIDIA and Kakao as supported partners. Apple support is coming soon but is not yet available.

The tool’s coverage is also its main practical limit. A positive result means the detector found a supported SynthID watermark, not that it can identify every kind of AI-generated media. A negative result does not prove a file was made by a person: systems outside the participating ecosystem may use another watermark or none at all.

Google DeepMind describes SynthID as an imperceptible digital watermark embedded directly in generated content. Google says the detector was initially available to journalists and other media professionals before access expanded to the public.

Google says it has applied SynthID watermarks to more than 180 billion images and videos and 240,000 years of audio. The announcement provided no accuracy, false-positive or false-negative rates, and gave no date for Apple support.