Artificial intelligence has turned political attack ads into a Wild West of fabrication. Candidates across the country are now flooding airwaves with doctored images and synthetic video of opponents in situations that never happened, from singing in dresses to shoveling money into furnaces to abandoning allies in foxholes.
The shift marks a fundamental break from campaign norms. What once seemed like dystopian fantasy is now routine. And there are almost no rules governing it.
Some campaigns volunteer to label their AI creations. Most do not. Democratic leaders have signaled they want to make disclosure mandatory if they win back Congress in the fall, but right now the practice sits in a regulatory void, blurring the boundary between documented reality and pure fiction in ways that reshape how voters see candidates.
Texas has become ground zero for the trend. A pro-Trump group called Citizens for Sanity recently released an attack ad depicting Democratic Senate nominee James Talarico in a dress, singing a twisted version of "Favorite Things" about transgender children. Talarico was not in that dress. He did not sing those words. The video was entirely synthetic.
The same candidate fell victim to an earlier AI attack from the National Republican Senatorial Committee in March. That spot showed him reciting old social media posts. The posts were real. Talarico's voice and image reading them were not.
The deepfake machinery is running in both directions. In Kentucky's 4th congressional district GOP primary, a "throuple" ad showed Rep. Thomas Massie checking into a hotel and holding hands with Reps. Ilhan Omar and Alexandria Ocasio-Cortez. Massie's own supporters fought back with AI footage of an elephant sporting Trump-like hair and a MAGA hat, and his challenger abandoning Trump in combat.
Georgia's gubernatorial race has devolved into parody. Brad Raffensperger used AI to show GOP primary rivals shooting wildly into the air and battling each other with pugil sticks. His competitor Burt Jones upped the ante with an entirely synthetic ad depicting his runoff opponent Rick Jackson feeding money into a furnace and blowing air into a hot air balloon.
Democrats are not innocent bystanders. Texas Democrat Jasmine Crockett inflated crowd sizes in one campaign spot using AI and posted synthetic video of herself as a baby alongside Trump and others. In New York City, Andrew Cuomo, running as an independent in the mayoral race, released AI footage of himself as a subway conductor, stockbroker, and window washer. Harry Dunn, a Democrat running for Congress in Maryland's 5th district, featured AI-generated men in suits labeling crypto and AIPAC while tossing golden basketballs in a carnival game.
The technology is getting easier and cheaper to deploy. Campaign budgets now routinely allocate resources to synthetic media production. Production timelines have compressed. The barrier to entry keeps falling.
What remains uncertain is whether voters can consistently spot the fakes or whether the sheer volume of synthetic content will eventually corrode the distinction between real and fake altogether. Early evidence suggests the latter risk is real.
Author James Rodriguez: "The AI genie is already out of the bottle in American politics, and there's no forcing it back without real regulation that neither party is rushing to impose when they're both benefiting from the chaos."
Comments