Google has a new tool to help detect AI-generated content

Google announced a caller SynthID Detector instrumentality astatine Google I/O that lets you cheque if contented has been made with the assistance of Google’s AI tools.

In a blog post, Google DeepMind’s Pushmeet Kohli describes SynthID Detector arsenic “a verification portal” that tin “quickly and efficiently place AI-generated contented made with Google AI.” It’s besides capable to “highlight which parts of the contented are much apt to person been watermarked with SynthID.”

SynthID watermarks are applied to AI-generated images, text, audio, and videos, including contented generated by Google’s Gemini, Imagen, Lyria, and Veo models, Kohli says.

Here’s however the instrumentality works, according to Kohli:

When you upload an image, audio track, video oregon portion of substance created utilizing Google’s AI tools, the portal volition scan the media for a SynthID watermark. If a watermark is detected, the portal volition item circumstantial portions of the contented astir apt to beryllium watermarked.

For audio, the portal pinpoints circumstantial segments wherever a SynthID watermark is detected, and for images, it indicates areas wherever a watermark is astir likely.

Google is starting to rotation retired the instrumentality to “early testers,” Kohli says successful the post. 

“Following the archetypal investigating phase, the portal volition gradually beryllium rolled retired to users who motion up to the waitlist to summation entree to the SynthID Detector,” Kohli tells The Verge. “We volition instrumentality learnings from this cohort of professionals and enactment to instrumentality contented transparency much broadly.”

I’m connected the waitlist, but I haven’t tested the instrumentality myself, truthful I can’t vouch for however good it mightiness work. And volition radical really usage it erstwhile it’s wide available? I anticipation so, but we’ll person to hold and see.