Midterm AI ad spending jumps to $137 million from $80 million

The Wesleyan Media Project now counts 220 AI-generated political ads with more than $137 million in spending behind them through October 11, 2026. Six weeks earlier, the same tracker put the total at just under $80 million.

The Wesleyan Media Project is a research group at Wesleyan University in Middletown, Connecticut that tracks political advertising in real time. It started coding ads for AI content in May 2026, and it has revised its running total upward twice since.

The project's October 15 release says the count covers ads "generated-at least in part-by using AI tools." The September 4 release described its own totals as looking at ads aired on broadcast television, connected TV and digital venues. The project calls the figure an undercount, because AI use is hard to detect and often undisclosed. "AI is making its way into the 2026 midterms in a big way this year," said Erika Franklin Fowler, a co-director of the project. "For comparison, we identified less than $50,000 in spending on AI-generated ads in the 2024 cycle."

How the AI ad count has moved

The project's own history shows how fast the category has grown. The May count found 33 ads worth about $20 million. The September 4 release raised that to 164 ads and just under $80 million. The new count adds 56 ads and roughly $57 million in six weeks, and the share carrying an AI disclaimer slipped from 31 percent to 27 percent over the same stretch.

Republican candidates and groups sponsored 81 percent of the identified ads and 88 percent of the spending, the project found. Democrats sponsored 12 percent, with independent candidates and nonpartisan issue groups accounting for the rest. The ads have appeared in 34 states.

Only 30 states require any AI disclaimer on political ads, according to Campaigns & Elections, and there is no federal rule. A disclaimer is one of the few signals the project's coders can catch, so undisclosed AI use goes uncounted.

Voters are not asking for more of it. An AAPC Foundation survey of 1,568 likely 2026 voters, conducted September 14 to 22, found 95 percent say it is unacceptable for a campaign to use AI to make a video of an opponent saying something they never said. Eighty-four percent object to an AI voice imitating a campaign's own candidate. Disclosure does not soften the view: 93 percent still object to the fabricated opponent video when it carries an AI notice.

On the professional side, a survey from the American Association of Political Consultants found 83 percent of political professionals use AI tools weekly or more, up from 59 percent the previous year, Campaigns & Elections reported.

The midterms are less than two weeks away, and the project updates its list as new ads air, so the $137 million will move again.

The full count, the coding method and every release are posted at the Wesleyan Media Project, where you can read the running total and the share of ads carrying a disclaimer.