Google makes SynthID accessible for verifying AI content

Google has opened access to its verification tool SynthID for the general public. Through the website synthid.com, users can check images, videos, and audio files for a digital watermark indicating whether the content is AI-generated. This feature supports material from both Google itself and technology partners such as OpenAI, Nvidia, and Kakao. A collaboration with Apple is in the works, but will be available later.

With SynthID technology, Google enables users to more easily investigate the provenance of images and other media. So far, Google has already marked 180 billion images and videos with watermarks, with over 240,000 years of audio material also equipped with this technology. The company receives about a million verification requests daily, highlighting how important this tool is for users who question the authenticity of media.

The operation of the verification tool is straightforward: SynthID applies a digital watermark to AI-generated media, and the tool specifically looks for this marking. This allows users to quickly check whether material comes from an AI tool. This can be particularly helpful for media disseminated through social platforms like WhatsApp. The approach aligns with developments by other companies such as Adobe and OpenAI, which are also working on invisible AI watermarks.

Although this verification feature offers accessibility for users, there are important caveats to consider. The SynthID tool only searches for its own watermarks; a file without this marking is not necessarily created by a human. It could also involve content produced by an AI tool that does not use SynthID technology. Therefore, the advice is to consider the results as supplementary information and to verify the original source and context of the file when in doubt.

Read the full article from AI Insider.