Google embeds watermarks in AI content

Every image and video from Google's Gemini models carries an invisible SynthID label in its pixels.

Checking for watermarks was clunky

Until now, the only way to check for SynthID outside Google's trusted tester program was to ask Gemini. The model's output could sometimes be confusing when you just wanted a simple yes or no.

Google launches SynthID.com detector

Starting today, anyone can visit SynthID.com to check content for AI watermarks. The previous detector only checked for Google's version of SynthID, even though other companies' images also carried the watermark.

Detector now supports multiple partners

Google's expanded detector supports distinct watermarks from OpenAI, Nvidia, Kakao, and soon Apple. Users can check a file for involvement of all those AI systems instead of accessing their detectors individually.

Login required with daily limits

You must log in using a Google, OpenAI, or Apple account to use the tool. Each user gets a daily usage quota of approximately ten image, video, and audio checks.

“a daily usage quota of approximately 10 image, video, and audio checks”

Limits prevent bypass tool development

Google engineers say limiting access is necessary to prevent people from using the checks to develop SynthID removal tools. The system may lock you out sooner if checking multiple very similar images.

System has significant gaps

Many AI model providers use different watermarking technology or no watermarks at all.Open-weight models run on personal hardware produce AI content with no watermarks, so plenty of AI content will not be identified by SynthID.