Unfiltered Deepfakes: How synthetic media is shaping the US midterm elections

A multimillion-dollar wave of AI-generated political advertising is spreading across US airwaves and social media ahead of the November 3 midterm elections, with both parties using hyper-realistic deepfake content against rivals. One video shows a Senate candidate laughing at cards reading "Death to America" and "Celebrate 9/11." Another targets Democratic candidate Abdul El-Sayed of Michigan, depicting him playing a board game with two people in kaffiyehs. It also shows a fabricated family photo of the Muslim candidate with four veiled women. "His family values are wrong for Michigan," the ad says, while a voiceover claims "destruction of Israel and the West" is "a mission" for them, according to a reported news agency.

Front Line Action, a political action committee aligned with Republican Senator Ted Cruz, was behind the El-Sayed video. According to a Federal Election Commission filing, it spent over $92,000 on advertising opposing him. The ad carries a disclaimer saying it was created at least partly with artificial intelligence, but researchers say most such ads do not.

The Wesleyan Media Project, which tracks election ad spending, has identified 220 political ads nationwide generated at least partly with AI tools. It estimates they cost more than $137 million, compared with less than $50,000 on AI-generated ads in the 2024 cycle. Republican candidates or groups sponsored 81 per cent of them, Democrats 12 per cent, and independents the remainder. Only 27 per cent carried an AI disclaimer. The group said its tally was almost certainly an undercount because it is hard to establish definitively whether AI was used.

"AI is making its way into the 2026 midterms in a big way this year," said Erika Franklin Fowler, the project's co-director, who called the lack of disclosure "problematic."

No federal rule requires political ads to carry an AI disclaimer, though some states have passed laws restricting AI use or requiring disclosure.

Democratic Senator Jon Ossoff told C-SPAN last month he found the use of deepfakes in political advertising "deeply troubling" and a "slippery slope toward the destruction of truth in political debate." He demanded removal of an undisclosed AI-generated ad that showed him receiving beauty treatments at a salon while covering up his "ugly record." The Senate Leadership Fund, a PAC dedicated to protecting and expanding the Republican Senate majority, was behind it.

The London-based Institute for Strategic Dialogue called the midterms the first US election in which AI is built into every major social media platform. The trend comes as public trust in the technology declines. In a survey by the nonprofit AAPC Foundation, 73 per cent of respondents called it unacceptable for campaigns to use AI-generated images or video in ads, and 95 per cent said it was unacceptable to create a video of an opponent saying something they never said.

Researchers, however, played down fears of an effect on results. New York University's Center on Tech Policy said in a report this month that such content is likely more effective at mobilising voters or raising money than at influencing candidate choice, and that its power to hold attention may be limited as audiences face more "AI slop."