Google Opens SynthID Detector to Public, Expands Partner Coverage

Google DeepMind expanded global access to SynthID Detector on Oct. 7, giving the public a way to check images, videos and audio for invisible AI watermarks across Google and participating partner systems.

The detector identifies supported SynthID signals embedded in media files. Current partners include OpenAI, NVIDIA and Kakao, while Apple is expected to join later.

SynthID marks are designed to remain detectable after common changes such as cropping, filtering, compression and speed adjustments. A missing SynthID signal does not determine how content was created.

Google has watermarked more than 180 billion images and videos and 240,000 years of audio. Verification features in Search, Gemini and Chrome process more than 1 million requests per day.

The public release follows the introduction of SynthID Detector in May 2025, when access was limited to journalists and media professionals testing the system. The tool now extends that verification process beyond professional users.

SynthID is designed to embed an imperceptible signal directly into AI-generated media. Google also supports C2PA Content Credentials, a separate provenance system that records how digital content was created or edited. SynthID provides a detectable watermark, while C2PA uses provenance metadata.

The system already appears in Google products that generate media. For example, Google’s Live Avatar feature for Gemini Live includes imperceptible SynthID watermarks in generated audio and video.

The public detector is available globally in English. The rollout gives users a way to check whether supported images, videos or audio contain a detectable SynthID mark, while the partner coverage continues to expand.