AI Attack Ads Flood 2026 Midterms as Disclosure Laws Lag

A 43-second video opens like a routine campaign spot: two figures introducing themselves as New York City Mayor Zohran Mamdani and Governor Kathy Hochul, strolling through a sunlit park. The conversation that follows casts both Democrats in a harsh light, with a voiceover calling Hochul "the political delusion pill for New Yorkers who would rather think about what four more years could actually cost them."

The video is not real. The images were generated with artificial intelligence, and the ad belongs to Bruce Blakeman, the Trump-endorsed Republican challenging Hochul. Small text noting it was "paid for by Blakeman for New York" runs throughout, but the use of AI is never flagged — despite a state law requiring such disclosure in political communications. The New York State Democratic Party filed a complaint with the state Board of Elections in August alleging the ads violate that law. Blakeman's campaign did not immediately respond to a request for comment.

Such ads are now a fixture of the 2026 midterms, ranging from the plausibly deceptive to the openly absurd. Republican Mike Rogers, locked in a tight Michigan Senate race against Democrat Abdul El-Sayed, posted a June video showing himself as a superhero halting a runaway car with one hand, captioned, "Haters will say it's AI." In North Carolina, an August ad from the campaign of Republican state Sen. Michael Lee depicts his Democratic opponent, yoga instructor Jessica Bichler, twisting into impossible poses — with a disclaimer reading, "You guessed it! AI was definitely used to generate these silly video clips."

The Wesleyan Media Project, which tracks election advertising, has counted at least 164 political ads generated or enhanced with AI this cycle, warning the true figure is likely higher because AI use is often hard to detect. Of those identified, 69% carried no disclosure. Republican candidates or pro-Republican groups sponsored 80%, pro-Democratic sponsors accounted for 11%, and the remaining 9% covered independents and issue-focused spots.

Cross-referencing the data with the National Conference of State Legislatures, project co-director Travis Ridout found that 32% of undisclosed ads aired in states with no AI disclosure laws still included a disclaimer, while only 29% of ads from states with regulations did. He cautioned that some ads fall outside these laws — Tennessee's statute, for instance, covers only deepfakes impersonating candidates — and that a law may have taken effect after an ad ran. Even so, voluntary disclosure surprised him. "It's a time of uncertainty and so it's all shaking out right now," he said, adding that norms may vary by consultant or firm.

Scott Babwah Brennen, director of New York University's Center on Technology Policy, said he is not opposed to labels on narrowly targeted, materially deceptive content, but argued the broader media environment must be addressed rather than individual cases. An October 2024 report he co-authored found that AI disclaimers can backfire: respondents rated candidates as less trustworthy and less appealing, and 37% did not recall seeing a label at all.

Democratic strategist Basil Smikle said missing disclosures may be intentional or simply an oversight, and noted that AI cuts costs and speeds production while risking authenticity and voter trust. "In the wrong hands, it could be a really dangerous trend, but if done well it can help do things that campaigns need to do and should do, which is tell good stories," he said.

With Congress and federal agencies yet to act, states have filled the gap — prompting calls for standardised rules. "If we could get something at the federal level I think that would make it clearer for campaigns what we can, what we can't do," Ridout said.