Deepfake Election News: Artificial Intelligence Voting Gaps

Two weeks before a Missouri primary, voters in St. Charles County started seeing a political ad that looked completely real. It wasn't. The whole thing was AI-generated, not a touched-up photo, not a filtered video, but a fully synthetic media production made by a local candidate in a county executive race. Not a Senate race. Not a presidential campaign. A county election.

Here's the thing. We've spent two years talking about deepfakes, AI-generated fake videos and images that look real, as a presidential-race problem. Famous politicians. Senate hearings. Big money. The assumption was that hyper-local races were somehow safe, too small to bother with. That assumption just died in St. Charles County.

The St. Charles County Ad That Changed Elections

The man behind the ad was Jason Law, a third candidate in the county executive race. He's been open about it: he saw a viral AI-made campaign ad from Spencer Pratt's Los Angeles mayoral run and thought, I can do that too. And he did. Cost? A fraction of what traditional video production runs. Time? Far less than hiring a crew.

That's the part nobody wants to say out loud. The price to produce a convincing AI political ad has collapsed. What used to require a studio, a budget, and a team now requires a laptop and an hour. Which means it's not just well-funded campaigns that can do this anymore, it's anyone running for anything, anywhere.

First Alert 4 broke this story, and the framing in coverage has been careful, nobody's saying Law committed a crime. Because in Missouri, he didn't. There is no law requiring him to label his ad as AI-generated. None. This article is part of a series, start with Retail Facial Recognition Washington Privacy Gap.

Deepfake Laws: Why Most States Lack Protections

As of May 2026, Biometric Update reports that 30 states have passed laws specifically addressing deepfakes in political communication, with Maryland being the most recent to join. Most of these laws don't ban AI content outright. They require disclosure: a label, a disclaimer, something that tells you "hey, this was made by a computer." That's it. A label.

Missouri isn't one of those 30 states. Illinois isn't either.

So if you live in one of those 20 states and a local campaign produces a fully synthetic attack ad about a school board candidate, a county commissioner, or a state representative, no label required. You scroll past it. Maybe you share it. Maybe your cousin argues about it in the family group chat. Nobody's legally obligated to tell you it was never real.

And at the federal level? There is currently no comprehensive law prohibiting or even regulating the use of AI deepfakes in political campaigns. Zero. Stack Cyber's state-by-state tracker shows the patchwork clearly: some states are serious, most are not, and the federal government is watching from the sidelines.

Deepfake Election News: Why Labels Aren't Enough

Even the states that have passed disclosure laws are running into trouble. California moved fast, AB 2839 was signed in September 2024, specifically targeting AI-generated political content. By August 2025, a federal judge had already struck down key parts of it, finding conflicts with Section 230 (the rule that protects platforms from being liable for what users post) and raising First Amendment concerns. Fast law, fast challenges.

But here's the deeper problem: even when a label exists, most of us miss it. Think about the last time you actually stopped to read the fine print at the bottom of a political ad. Nobody does. We see the image, we register the message, and we move on. The label is compliance theater, it checks a legal box without changing what actually happens in a voter's brain at 9pm on their phone. Previously in this series: Cheshire Police Live Facial Recognition Operation Vigilant P.

"Most voters do not know the difference between what is real and what is AI-generated, placing a greater responsibility on the public, and without any laws regulating AI, it places a greater burden on the voter, on the citizen to be informed." Expert commentary via First Alert 4

That quote lands differently when you picture who's actually watching a county executive race ad. It's not a political journalist with a fact-checking team. It's a parent who just finished dinner, half-watching their phone, trying to remember whether to vote Tuesday or Wednesday. (Tuesday, always Tuesday.) They're not running a media literacy seminar in their head. They're tired.

Your Brain Is Already Being Played

There's a psychology concept worth knowing here, the availability heuristic (big words, simple idea: your brain treats "easy to picture" as "probably true"). When you see a video of something, even if it's fake, your brain files it as a memory. Later, when someone asks if a candidate did or said something, your brain pulls up that file. The label you maybe-saw doesn't come with it.

A 2025 study published in the Journal of Creative Communications confirmed what many suspected: people genuinely cannot reliably tell AI-generated video from real footage. This wasn't a study of tech-unsavvy seniors, the detection failure was widespread. As TrueScreen's research summary puts it, the technology to generate synthetic media is running years ahead of the technology to detect it. That gap is where campaigns live right now.

"Typically, as is the pattern in American politics, when some new technology comes on the scene, that technology tends to outpace the law. But unlike previous tech gaps, deepfakes skip the verification stage entirely, voters no longer know what to distrust." Expert analysis, Washington University School of Law AI Policy Resources

"Voters no longer know what to distrust." Read that again. It's not that people are gullible. It's that the mental shortcut we all use, I'll know a fake when I see onehas quietly stopped working. And the people running local campaigns now know it.

What You Can Actually Do About It Right Now

Look, nobody expects you to become a media forensics expert before the next primary. But there are two things worth doing today. Up next: Your Face Is Being Scanned At The Grocery Store And Washingt.

First: before you share any political video from a local race, ask one question, where did this come from? Not "does it look real," because as we've established, that test no longer works. But "did this come from a verified campaign account or an official outlet?" Synthetic content almost always enters your feed through a shared post, not a primary source. If you can't trace it back to the candidate's official page or a named news organization, treat it like you'd treat a forwarded WhatsApp rumor.

Second: if you've ever looked at a photo or a video and wondered whether the person in it is actually who they claim to be, that instinct is exactly right. The question of "is this face, this voice, this image actually real?" is no longer paranoia. It's basic digital literacy in 2026. Tools that verify the authenticity of images and identity documents exist precisely because this problem is here now, not coming someday.

The Axios reporting on the national surge of AI ads across multiple 2026 campaigns makes clear this isn't a Missouri quirk, it's a preview. The question now isn't whether this happens in your state. It's whether your state will tell you when it does.

Jason Law said he was inspired by a Los Angeles mayoral candidate's viral AI ad. Someone in your state is watching the St. Charles story right now, thinking the exact same thing he did: I can do that too. The only real question is whether they'll bother with a label, and in most of the country, they're not legally required to answer it.

Deep Fakes and the Speed of the 2026 Campaign Cycle

The deep fake problem in this election cycle is not a future risk, it is already running in live primaries. A deep fake ad can be built, tested, and posted within a single afternoon, which is faster than any newsroom or election office can typically respond. That speed is exactly why the St. Charles County case matters: it shows how quickly a deep fake can move from a candidate's laptop into thousands of voters' feeds before anyone official even knows it exists.

Why Fakes Spread Faster Than Corrections

Fakes travel through share buttons, not fact-checking desks, which means they often reach far more people than any correction ever will. Once a fake political video is shared into a local Facebook group or group text, the platform's algorithm treats it like any other engaging post and keeps pushing it outward. By the time a news outlet or election office confirms the video is fake, most of the damage to voter perception has already happened. This is part of why disclosure laws focus on stopping fakes before they publish, not cleaning them up after.

Deepfake Disinformation Aimed at Local Ballots

Deepfake disinformation used to be described mainly as a national security worry tied to foreign interference in presidential races. The St. Charles County ad shows that deepfake disinformation now works just as well, arguably better, in small, low-turnout local elections where a single ad can shift the outcome. Because local races draw so little outside scrutiny, deepfake disinformation aimed at a school board or county commissioner race can spread almost entirely unchecked.

Political Deepfakes Are Now a Down-Ballot Tool

Political deepfakes were once treated as something only a national campaign with a big budget would attempt. The collapse in production cost means political deepfakes are now within reach of a single candidate, a small PAC, or even an angry volunteer with a laptop. That shift moves political deepfakes from a rare headline event into a routine campaign tactic that voters should expect to see in almost any contested race.

What Political Campaigns Gain From Going Synthetic

Political campaigns that use AI-generated video get a message that looks polished without paying for a production crew, actors, or studio time. For political campaigns operating on a shoestring budget, that tradeoff is obviously appealing, even if it raises serious questions about honesty with voters. As more political campaigns see how cheaply this content can be made, the incentive to skip disclosure, where no law requires it, only grows stronger.

Deepfakes as a Normal Part of Election Advertising

Deepfakes are shifting from a novelty story to a standard line item in how some campaigns plan their advertising. Once one candidate in a race uses deepfakes and gains an edge, rivals in the same election often feel pressure to respond in kind rather than get left behind. That competitive pressure is one reason deepfakes are likely to show up in more local races before more states pass disclosure requirements.

Election Officials Face a New Kind of Problem

An election office is built to manage ballots, polling places, and vote counts, not to referee whether a viral video is real. When an election dispute involves a synthetic video instead of a paperwork error, most local election staff have no clear protocol to follow. That gap means a contested election claim tied to AI content can linger unresolved long after the votes are actually counted.

Elections Without a National Disclosure Standard

Elections across the country are currently governed by a patchwork rather than one clear federal rule on AI-generated political content. That means the same kind of ad can be legal in one state's elections and against the law in the next state over. Until Congress or the courts settle a single standard, elections will keep producing these state-by-state gaps that campaigns like Jason Law's can legally use.

Artificial intelligence is the plain term for the technology behind every one of these ads, and it helps to say it directly instead of only using shorthand like "deepfake" or "synthetic media." When a local election story mentions artificial intelligence, it usually means software that was trained on real photos, video, and audio, then used to generate a new clip that never actually happened. That's the whole story in one sentence: artificial intelligence made a video of an event that never occurred, and in most states nobody has to tell you that.

Voting is the part of this story that never changes even as the technology around it does. You still walk into a polling place, or mail in a ballot, and make a choice based on what you believe about the candidates. The problem the St. Charles County ad exposes is that the information feeding into your voting decision can now be manufactured, cheaply, by anyone with a laptop, and in most states there's no requirement to flag it before you vote.

Election security conversations have traditionally focused on protecting voting machines, ballot counts, and voter rolls from hacking or tampering. The St. Charles County case is a reminder that election security also has to cover the information voters see before they ever reach the ballot box, since a convincing fake ad doesn't need to touch a single voting machine to change an outcome. As more local races see AI-generated ads, election security discussions will likely expand to include content authenticity, not just infrastructure protection.

Election disinformation isn't a new problem, but the deepfake version of it moves faster and looks more convincing than the doctored photos and misleading captions of years past. What makes today's election disinformation different is production speed: a fake ad that once took a skilled team days to fake convincingly can now be generated by one person in an afternoon. That shift is exactly why the gap between the 30 states with disclosure laws and the 20 without matters so much for how election disinformation spreads.

Fake disinformation built from synthetic video puts local election officials in a strange spot, since their job has never included fact-checking viral ads. When fake disinformation shows up in a county race, there's often no clear person whose job it is to confirm or debunk it before voters go to the polls. That gap between when fake disinformation appears and when anyone officially responds is where most of the damage happens.

Deepfake technology keeps getting cheaper and easier to use, which is the single biggest reason this story moved from presidential races down to a county executive contest in Missouri. A few years ago, deepfake technology capable of producing a believable political ad required real technical skill and expensive software. Today, deepfake technology is packaged into consumer-friendly tools that a candidate can learn in an afternoon, which is exactly what happened in St. Charles County.

An election deepfake doesn't need to be perfect to work, it just needs to be convincing enough to hold up during a quick scroll through a social feed. Most voters aren't studying a video frame by frame looking for glitches; they're deciding in seconds whether to believe what they just saw. That's why an election deepfake aimed at a low-information down-ballot race can be more effective than one aimed at a heavily scrutinized national campaign.

Deep fake detection tools exist, but they're built mostly for researchers and journalists, not for the average voter scrolling through a group chat. A deep fake that fools a casual viewer for even a few seconds has usually already done its job, since the emotional impression sticks even if someone later learns the video wasn't real. Closing that detection gap for ordinary voters, not just professionals, is one of the bigger unsolved problems this story points to.

Deep, in the context of this story, refers to the "deep learning" techniques that let software study thousands of real images or audio clips and then generate new, convincing fakes. That deep technical process is invisible to the viewer, all you see is the finished video, not the machine learning underneath it. Understanding that deep learning is the engine behind these ads helps explain why they keep getting more realistic every election cycle.