Google Opens SynthID.com Detector for 180 Billion AI Files as Support Expands to All Partners

3 articles · Updated · Ars Technica · Oct 7

Google on Tuesday made its SynthID detector publicly available at SynthID.com, letting anyone check images, video and audio for hidden AI watermarks without going through Gemini.
The expanded tool now recognizes distinct SynthID implementations from all partners, including OpenAI, Nvidia and Kakao, with Apple support coming soon—fixing a key gap in Google's earlier detector.
Google said Gemini models alone have embedded SynthID in more than 180 billion images and videos and 240,000 years of audio, making broader verification more useful as adoption grows.
The public site gives a simpler yes-or-no result on whether media was produced or edited with SynthID, rather than the internal tool's more detailed region-level highlighting.

Sources

Left50%

Center50%

If SynthID only flags participating AI models, are we leaving the door wide open for untraceable deepfakes from rogue developers?

Why did Google release a simplified public AI detector while keeping its most powerful forensic tracking tools locked away internally?

Could heavy image editing strip away Google's invisible AI watermark, giving malicious actors a perfect loophole to spread misinformation?