AI-Made Campaign Ads Top $187 Million Ahead of the Midterms, Study Finds

Campaigns and political groups have spent more than $187 million airing AI-generated or AI-assisted ads in the 2026 election cycle, according to the Wesleyan Media Project. The group, which tracks political advertising, counted at least 289 such ads as of Oct. 29.

That is a sharp rise from 2024, when the group found barely $50,000 in AI ad spending for the whole cycle. “It’s kind of gone from zero to 100 in just that two year term,” said Michael Franz, a Bowdoin College government professor who co-directs the project.

In its Oct. 15 update, the project had counted 220 ads and $137 million. The total has grown since, and Franz said the count is probably still low. It misses more targeted ads on social media, and AI use is hard to detect and not always disclosed.

Scenes that never happened

Most of the ads identified so far fall into two groups, Franz said. Some use AI to polish a spot, such as replacing stock footage. Others are obvious fakes played for laughs, a format researchers call “slopaganda.”

One ad from the campaign of Texas Attorney General Ken Paxton, who is running for Senate, shows his opponent, state Representative James Talarico, marking up a Bible with a crayon and talking about religion and gender. A spot from Nassau County Executive Bruce Blakeman, running for governor of New York, shows Governor Kathy Hochul gardening with New York City Mayor Zohran Mamdani in an ad styled like a drug commercial. An ad from former Representative Mike Rogers, running for Senate in Michigan, shows his opponent Abdul El-Sayed grinning inside a Trojan horse under a banner reading “Definitely NOT Communism.” None of those scenes happened.

Some ads do say AI was used. A spot targeting Jessica Bichler, a Democratic candidate for state Senate in North Carolina, says its clips of Bichler stretching were generated with AI. But of the ads counted through mid-October, only 27 percent carried a disclaimer.

Who is spending

Republican candidates, a Republican Party committee, or Republican-leaning groups account for just over 90 percent of the spending on AI-generated or AI-enhanced ads through late October, the project found.

Democrats use the tools as well. In August, the campaign of California state Senator Scott Wiener, who is running for the House seat held by Nancy Pelosi, built an AI chatbot lampooning his opponent, San Francisco Supervisor Connie Chan. Billboards sent voters to ConnieChan.ai, where the bot introduced itself as the first AI trained to say no to anything. Pelosi and several other female elected officials objected in a letter. “Women have spent generations fighting to have our own voices heard in the halls of power,” they wrote. “We should not have to fight against artificial intelligence putting words in our mouths.”

Talarico, who has been targeted repeatedly, told CNN in September that “Texans are sick of these AI ads trying to deceive voters.” Candidates facing complaints about disclosure, including Blakeman, have argued some spots are exempt because they are satire.

A patchwork of rules

No federal law governs deepfake campaign content. In 2024, 31 states adopted resolutions or enacted legislation, most of them requiring disclosure, according to the National Conference of State Legislatures.

Voters head to the polls Tuesday, Nov. 3. AI is also reaching them outside ads. Four percent of likely voters said in September they had consulted AI for election information in the previous month, according to a survey by the American Association of Political Consultants Foundation. Thirty-eight percent said they had gotten such information from AI search summaries, half of them without realizing it.

What effect the ads have on votes is unclear, researchers said. “It’s really hard to persuade voters of anything,” said Isabel Linzer, an elections and democracy analyst at the Center for Democracy and Technology. The longer-term worry, she said, is that a flood of fakes could leave voters unsure whether anything they see is real.