Google launches SynthID Detector for AI-generated media

Google has launched the SynthID Detector to the public, a tool designed to identify AI-generated images, videos, and audio files. The tool works by searching for imperceptible digital watermarks embedded within the media by specific AI models.

While originally focused on Google’s own technology, the detector now supports content created by OpenAI, NVIDIA, and Kakao, with support for Apple’s AI-generated media expected to be added soon. Users can access the service via synthid.com by logging in with a Google, Apple, or ChatGPT account and uploading a file for analysis.

Google noted that since the launch of SynthID in 2023, it has applied watermarks to over 180 billion images and videos, as well as audio content totaling an equivalent of 240,000 years. It is important to note that the tool is not a universal AI detector; it specifically searches for compatible SynthID watermarks. If a piece of media was generated by an AI model that does not use this specific watermarking technology, the detector will not be able to identify it as artificial.