In brief
The 2026 U.S. midterm elections are seeing a dramatic rise in AI‑generated political advertisements, with over $137 million spent on at least 220 ads identified by the Wesleyan Media Project. Experts caution that the lack of disclosure could mislead voters and undermine electoral integrity, while tech lobbyists push back against tighter regulation. Disinformation content ranges from provocative visuals to extremist messaging, highlighting challenges for moderation and transparency.
As U.S. voters headed to the November 3 vote, an influx of AI‑driven political ads flooded social media feeds, experts say, raising alarms about misinformation and electoral integrity.
The Wesleyan Media Project, which tracks the use of AI in election advertising, identified roughly 220 AI‑generated ads worth more than $137 million in October alone, a sharp jump from the $50,000 spent on AI ads during the 2024 cycle. The report described the figure as an “almost certainly” low estimate, citing the difficulty of detecting AI‑generated content.
Many of the ads carry no disclosure that AI was used. Illustrative examples include a politician‑shaped image of a Republican gubernatorial candidate embracing former President Clinton under the slogan “Clinton Lover. Trump Hater.” Another depicts a Democratic Senate candidate laughing at cards reading “Death to America” and “Celebrate 9/11.” One more shows a Florida Republican primary candidate hosting a party with Iranian leaders brandishing scimitars that target the Star of David, a symbol of Judaism.
Those who run misinformation campaigns can evade moderation when they “keep tweaking the post a little bit each time you create them, and some of them are going to get through,” Frances Haugen, a former Facebook data engineer and whistleblower, said. She added, “You can run an information operation now with quality that was impossible in 2016 for super, super cheap.”
An American Association of Political Consultants Foundation survey found only 1% of respondents considered AI their most credible source of election information, yet 20% of voters plan to use AI to research a race on their ballot, and twice that number has already received election details through an AI search summary.
Co‑opting AI, a cottage industry has sprung up, advising candidates on shaping what AI says about them. The left‑wing advocacy group Run for Something launched CampSight, a platform that tracks newly surfaced AI content across the political spectrum.
Amid the rise in AI‑generated political material, tech firms have intensified lobbying against stricter AI regulation, asserting the economic importance of their platforms while regulators face pressure to tighten disclosure rules.
While the 137 million dollars spent on AI ads in October is the highest figure publicly reported, analysts emphasize that it is most likely an undercount. Detecting AI usage remains technically challenging, and the lack of disclosure hampers public scrutiny of who funds and creates these messages.
Despite advances in moderation techniques, the industry’s capacity to insert tiny variations into each ad suggests many will slip past current safeguards, leaving voters uncertain about the authenticity of the content they see.
Uncertainties remain: the exact scale of AI‑generated advertising beyond October data, the effectiveness of existing content‑moderation strategies, and the full impact of undisclosed AI propaganda on voter understanding and electoral outcomes are still unresolved.
