SynthID crosses 100B marks and spreads to OpenAI, Kakao

Watermarking becomes a cross-platform trust layer

Google DeepMind’s May 26 post turned SynthID from a model feature into an ecosystem story. The tweet said SynthID has watermarked “over 100 billion pieces of content” and named OpenAI, ElevenLabs, and Kakao as partners adding the technology to their models. The source tweet is available on X.

The linked Google post gives the bigger picture. SynthID embeds imperceptible signals into AI-generated media, and Google says it has now been applied to more than 100 billion images and videos plus 60,000 years of audio. Verification is also moving outward. Google says SynthID verification in Gemini has been used 50 million times globally, and that the feature is expanding to Search now and Chrome over the coming weeks.

The partner list matters because watermarking only works at web scale if outputs from many generators carry compatible signals. Google says more content from OpenAI, Kakao, and ElevenLabs will carry SynthID, building on previous work such as open-sourcing SynthID text watermarking and partnering with NVIDIA on AI-generated video from Cosmos world foundation models. That makes provenance less dependent on a single app and more likely to show up in the places where people encounter media.

Google is also pairing SynthID with C2PA Content Credentials. Pixel 10 already supports credentials for images in the native camera app, and Google says Pixel 8, 9, and 10 phones will add video support in the coming weeks. The next test is policy, not just detection accuracy: whether search engines, browsers, social platforms, and enterprise review systems treat these signals consistently when labeling AI media, fighting fraud, or preserving proof of origin.