Google has expanded SynthID Detector from a limited media-professional tool to a public service available globally in English. Users can upload images, video or audio and check for the imperceptible SynthID watermark used by Google AI systems and participating partners including OpenAI, NVIDIA and Kakao, with Apple support announced as coming soon.
SynthID does not attempt to infer from visual style whether arbitrary media is AI-generated. Instead, it detects a watermark embedded during generation. That distinction is important: finding the watermark provides provenance evidence, but failing to find one does not prove that a file was created by a human or that it was never edited with AI.
Google says it has watermarked more than 180 billion images and videos and an amount of audio equivalent to 240,000 years. It also says built-in verification features across Search, Gemini and Chrome now handle more than one million requests per day. These scale figures are Google-reported.
Independent coverage from Ars Technica, UOL and others confirms the public rollout and highlights the same limitation: SynthID Detector is useful for supported watermark ecosystems but is not a universal detector for all synthetic media. The launch is therefore best understood as a provenance product and an interoperability expansion rather than a complete solution to AI-content authenticity.