That Attack Ad About Your Local Candidate? A Laptop Made It, and No One Has to Tell You.

Full Episode Transcript

A candidate for county executive in St. Charles, Missouri, paid for a political attack ad. The people in the video? They weren't filmed. They were generated by a computer. And under Missouri law, nobody has to tell you that.

Trusted by Investigators Worldwide

Run Forensic-Grade Comparisons in Seconds

Detailed facial comparison reports. Results in seconds.

Get Started

7-day refund guarantee**

If you've ever scrolled past a political ad on your

If you've ever scrolled past a political ad on your phone, this story is about you. Because the technology that used to be reserved for big Senate scandals just landed on a county ballot. A candidate named Jason Law paid for a fully synthetic ad in the St. Charles County executive race. He said he got the idea from a viral A.I. ad in a Los Angeles mayoral campaign. The vote was about two weeks away when it started drawing scrutiny. So the question threading through everything today is simple. When a fake video costs almost nothing to make, who protects the voter?

Start with the money. For years, convincing fake video was expensive and slow. Now a single county-level candidate can afford it. That's the real shift. If one challenger can produce synthetic political video, so can local groups and rivals everywhere. The cost barrier didn't lower. It basically collapsed. For your local race, that means the next attack ad about a school board seat or a county judge could be built on a laptop overnight.

Now the law, or the lack of it. As of May of last year, thirty states had passed laws addressing deepfakes in political ads. Maryland was the thirtieth. But Missouri isn't one of them. Neither is Illinois. That leaves voters in about twenty states with no legal right to a disclaimer. And there's no federal law that bans deepfakes of real people in campaigns. So whether you get warned depends entirely on your zip code.

Here's the part that should stop you. A study published in the Journal of Creative Communications last year tested whether people could spot deepfake videos. They couldn't. A separate pre-registered experiment reached the same conclusion, people do not reliably tell real from fake. That matters because most of the new laws only require disclosure. They ask campaigns to label the fake, not to stop making it. But a disclaimer only helps if you notice it. Picture a tired voter scrolling ten days before a primary. They remember the ad. They never see the fine print.

The Bottom Line

And even the tough laws are shaky. California passed one in the fall of two years ago. By last summer, a federal judge struck down parts of it. The judge found it likely violated free speech protections. That changes how lawmakers write these rules going forward. For the rest of us, it means the safety net you assume exists might not hold.

Here's the twist most people miss. Every one of these laws assumes someone, a platform, an official, a voter, can actually identify a fake. But the research says we can't. The real problem isn't weak regulation. It's that the detection technology to back it up doesn't fully exist yet.

So let's make it plain. A local candidate ran a fake video ad, and in most states, no one has to warn you it's fake. The tools to make these ads are cheap now, and the tools to catch them are still catching up. The old rule was "seeing is believing." That rule just expired, for a Senate race and for your county ballot alike. Before you share that outrageous clip about a local candidate, it's worth asking who really made it. The full story's in the description if you want the deep dive.

That Attack Ad About Your Local Candidate? A Laptop Made It, and No One Has to Tell You.

Do political ads have to say if they use AI-generated video?