Google has opened SynthID Detector, its AI watermark checker, to the public after limiting early access to journalists.
The public release follows more than a year of work on the website at synthid.com, which scans content for SynthID watermarks, the invisible markers that Google’s AI models and those of its partners embed in media.
Google already reads those watermarks, along with C2PA content credentials, through Circle to Search and the Gemini app, but the standalone website opens the capability to people outside Google’s AI ecosystem.
Which AI tools the detector recognises
Beyond content from Google’s own models, the detector recognises images, video and audio clips that AI tools from Kakao, NVIDIA and OpenAI produce, which widens its reach across rival generative AI platforms.
Apple-generated images come next, as Google plans to add support soon for pictures that Apple Intelligence features such as Image Playground create or edit, which would make SynthID Detector one of the first checkers to handle Apple’s output.
Where watermark detection falls short
Even with Apple on board, the website only recognises content that carries a SynthID watermark, which leaves gaps wherever AI companies rely on different watermarking standards or where someone strips the markers from a file.
Meta faces the same fragmentation with its own experimental web tool, which identifies images from its latest Muse Image models but cannot read output from other companies or its own earlier models.
That patchwork means no single service works as a universal checker, so anyone verifying a suspicious image or clip may need several tools depending on which company’s model produced it.
Google’s results also stop short of explaining how a piece of media came together or which parts are synthetic, since the tool makes no distinction between fully AI-generated content and material that AI only edited.
Despite those limits, access to the website costs nothing, although each visitor must sign in with a Google, OpenAI or Apple account before running a check on any image, video or audio clip.
Google plans to widen the detector’s coverage further, since Apple Intelligence support arrives soon, ElevenLabs has committed to adding SynthID watermarks, and Google’s website notes ongoing work to extend detection partnerships to more generated media.
The post Google’s AI detection website is now live appeared first on Trusted Reviews.