Google yesterday launched SynthID Detector, an online tool that “helps you make informed decisions about what to trust” on the internet. It does this by identifying AI-generated content, whether that’s an image, video, or audio file. If Google needs a special tool to help people identify one of the internet’s major features, maybe that feature isn’t as useful as it’s being made out to be.
The tool isn’t new. It’s been around, assisting “media professionals [in verifying] AI-generated content,” since last year. It’s just the broader rollout to the general population that’s new. It’s also not a catch-all solution, since it’ll only spot content with a SynthID identifier embedded.
SynthID Detector
That means it’ll tell users if something was made with one of Google, Nvidia, OpenAI, or Kakao’s generative platforms. If it’s sourced from somewhere else, it’s not going to be capable. See, those platforms include a SynthID watermark when creating something. It’s this bit that SynthID Detector (as the name suggests) detects. According to Google, “we’ve watermarked over 180 billion images and videos, along with 240,000 years of audio content.”
That’s… a lot of stuff that needs checking. The tool will also eventually incorporate content from Apple. It doesn’t seem like the iPhone maker will add a ton of content to the pile, but you never know.
But if you’re planning to use it yourself, Google will make you promise not to steal its ideas, circumvent its systems, or make multiple accounts to cost the company a ton of money by running large volumes of queries. You’ll also need to be signed into the tool, either with a Google, Apple, or ChatGPT login, before it’ll begin working on your behalf.
SynthID Detector is supposed to help users “navigate the web with confidence,” according to Google. That’s one heck of an admission from a company that thinks AI is incredibly awesome at… well, everything.
