Google opened SynthID Detector to the public globally in English on October 7, 2026. The tool checks images, video and audio for compatible SynthID watermarks—a narrower task than identifying every kind of AI-generated media.
What a SynthID result tells you
SynthID is Google’s watermarking technology. Google DeepMind describes its marks as imperceptible to people and designed to be detected in generated media. A positive result indicates that the Detector found a compatible mark. If it finds no mark, that does not prove a file is authentic or human-made; the check concerns SynthID watermarks, not AI authorship in general.
Which media and partners Google named
Google announced checks for images, video and audio. It named OpenAI, NVIDIA and Kakao among the partners whose media the Detector can check. Google described Apple as a forthcoming addition.
SynthID technology also covers text watermarking, but Google’s announcement of the public Detector listed image, video and audio checks.
The scale Google reported
As of October 7, 2026, Google said SynthID marks had been added to more than 180 billion images and videos and 240,000 years of audio. The figures are cumulative totals for different kinds of media.
Google also said verification features integrated into Search, Gemini and Chrome handled more than one million requests a day. That figure covers those integrated features, not the standalone Detector.
One hands-on demonstration
An English-language walkthrough shows one ChatGPT-generated image receiving a result that it was made or edited with OpenAI. In a separate test, a video returned “SynthID not detected.” These are individual observations, not a measure of the Detector’s overall accuracy.
