Google has launched a web-based SynthID Detector tool designed to identify images, video, and audio created or edited by AI models. The service is now available to the public, though users must sign in with a Google, OpenAI, or Apple account to access it.
The tool currently supports content generated by models from Google, OpenAI, Nvidia, and Kakao. Google plans to expand these capabilities in the coming weeks to include images produced by Apple’s AI tools, such as Image Playground.
SynthID, originally developed by Google in 2023, functions as an imperceptible watermark embedded in AI-generated media. Previously, access to detection tools was limited primarily to journalists and media professionals.
The detector has significant limitations, as it only identifies content containing a SynthID watermark. It fails to detect AI media from providers like Anthropic or xAI, as well as thousands of open-weights models from China that do not use this standard.
Furthermore, the tool cannot differentiate between entirely AI-generated content and media that has only been edited with AI. It is unable to isolate specific fake elements within a file, highlighting the lack of a universal industry standard for watermarking.
