We’ve run a number of stories on the problem of AI-generated deepfakes — the problems they create, the panic they generate, and the legislation that can result.

In the New Scientist video at right, we are told that one solution is more government regulation of Big Tech companies. While European legislators may like that strategy, it is not practical. The tech bros can ignore and outwit them.

At Gizmodo, Matt Novak reports on a new watermark-based detection system from Google that might help.

For example, in the last few days, a number of versions of this sort of photo (below) of tech titans appeared on X and BlueSky. A thoughtful viewer would immediately suspect deepfake, of course. What are they all doing together in a parking lot? But doubt is not the same thing as knowing that something is faked.

From Novak:

Right off the bat, it should be said that the vast majority of AI image detectors are not reliable. Many people think you can use tools that are openly available on the web and figure out if a given image is AI. But they’re not good. For example, people often ask Grok on X whether a photo was created with generative artificial intelligence. And it frequently gets the answer wrong. Sometimes in amusing ways.

Google developed an AI watermark called SynthID a couple of years ago, but the company didn’t allow the average user to check whether an image had the watermark. That changed just a few days ago. Now anyone can upload an image to Gemini and ask if it has the SynthID watermark, which is invisible to the naked eye.

“Those Viral Photos of Elon and Zuck Are AI. But Google Launched a New Way to Check for Fakes,” November 25, 2025

All images created with Google tools will have the watermark. From Google: “If you see an image and want to confirm it has been made by Google AI, upload it to the Gemini app and ask a question such as: “Was this created with Google AI?” or “Is this AI-generated?” Gemini will check for the SynthID watermark and use its own reasoning to return a response that gives you more context about the content you encounter online.”

Needless to say, the cigar-chomping “Musk” and his tech bros — in their many iterations — were quickly outed as fake.

But there is a serious side to all this. False evidence, introduced in court, might be similarly shown up.

There will likely be pressure on other image tool creators to embed watermarks and, with any luck, suspicion will fall on those that don’t.