Google opened its SynthID Detector to the public this week, putting a free portal at synthid.com where anyone can upload an image, video, audio clip or text sample and check whether it carries an invisible AI watermark. The October 7 launch expands what began as a limited tool into a global, English-language service.
One SynthID Detector, Rival Labs at the Table
The guest list is the story: OpenAI, NVIDIA and Kakao are participating in the detector network, with Apple expected to join. In practice, content watermarked under different labs’ systems can be checked in one place – closer to a shared bus lane than a walled garden. The timing also lands amid the EU AI Act’s transparency push, which wants AI-generated content identifiable, and watermarks are the least intrusive tool in that toolbox.
The Scale of the Paper Trail
Google says more than 180 billion images and videos and roughly 240,000 years of audio have already been watermarked with SynthID. Detection is not foolproof – determined re-encoders can strip marks – but a public, cross-vendor check makes ‘I had no idea it was AI’ a slightly weaker defense.
OpenAI, for its part, is separately rolling out its own invisible ‘textGrain’ watermarking for ChatGPT and Codex output in the EU. Two labs, two brands of invisible ink, one direction of travel: the era of the unmarked AI artifact is winding down, and the detector anyone can use is now one URL away.
Source: Google blog and TechCrunch (October 7, 2026).
