AI reshapes campaign outreach as election officials warn of deepfakes

The Federal Election Commission voted Sept. 19, 2024, not to open a new rulemaking on AI in campaign ads and instead adopted an interpretive rule saying existing fraudulent-misrepresentation rules already cover deceptive AI-generated campaign content. The decision came after a July 2023 petition pressed the commission to make that point explicit, as campaigns were already folding generative AI into the machinery of modern politics.

Behind the most visible AI images, campaigns and ad firms used the technology to analyze voter data, write first-draft messages and produce fundraising solicitations at scale. Brennan Center researchers said political campaigns were already using these tools for ad copy and donor asks, while a Center for Media Engagement analysis found campaign uses that ranged from virtual candidate avatars and phone outreach to routine fundraising appeals and first-pass opposition research. In practice, that made AI less a novelty than a force multiplier for the data-driven tactics that already define U.S. politics.

The more consequential shift was in one-to-one persuasion. A 2024 article in LSE Public Policy Review warned that when campaigns can combine AI with personal contact data, AI-to-voter communication could become especially disruptive and alter the political playing field. That concern fit an election cycle in which campaigns were under pressure to move faster, personalize more aggressively and test more messages than human staff could manage alone.

The integrity debate ran in parallel. The Associated Press described 2024 as a moment when distrust and disinformation were challenging the basics of how America votes, and AI made false audio, images and video cheaper to produce and easier to spread. One AP report said an AI deepfake was believed to be among the first examples of a deepfake deployed in a U.S. political race, sharpening fears that synthetic media could reach voters before fact-checkers or regulators could respond.

Not all worst-case warnings materialized in full, but the threats proved real enough to shape the race. AP VoteCast surveyed more than 120,000 voters across all 50 states from Oct. 28 to Nov. 5, 2024, capturing an electorate voting amid broad distrust and heavy digital targeting. A Reuters Institute for the Study of Journalism analysis also found that public AI detectors tested in February 2024 had limits in spotting deepfakes, underscoring how difficult it remained for voters, campaigns and election officials to tell synthetic content from the real thing.