American voters are suspicious of artificial intelligence, and American campaigns are largely unmoved by their suspicion.
A disconnect at the heart of modern campaigning
A Pew Research Center survey of more than 5,000 adults this year found that two-thirds of Americans lack confidence in the government's ability to regulate AI responsibly, a majority think it is advancing too fast, and many expect it to do more harm than good over the next two decades. A separate Pew survey found that about seven in ten would think less of a candidate upon learning AI had written a speech they admired.
None of that has slowed the industry down.
What campaigns are actually doing
AI has moved well past novelty in the 2026 midterm cycle. Consultants in both parties use generative tools to produce first drafts of fundraising appeals, social-media copy, opposition-research summaries and canvasser scripts — what one strategist called "personalization at scale." Republican operatives have moved especially aggressively, Axios reported, using AI to model voter sentiment around live news events and monitor social media in near-real time; Democratic strategists have generally been more cautious, citing privacy and job-displacement concerns.
For targeting, the parties lean on established voter-data platforms now layered with machine learning — and newer tools that send AI-generated text messages trained on a candidate's own words, capable of holding thousands of automated "conversations" with voters in a single race.
A new era of attack ads — and disclosure fights
The most contentious frontier is video. AI-generated political ads — some clearly labeled, some not — have proliferated this cycle, and fabricated "influencer" accounts have surfaced on social platforms, raising fears about deepfakes and undisclosed synthetic media.
The law has struggled to keep up. Roughly 30 states have enacted some form of rule on AI in political advertising, most requiring disclosure when an ad uses AI-generated or substantially altered imagery. But enforcement is uneven, and a federal judge struck down key parts of California's law in 2025 on First Amendment grounds. No national disclosure standard exists, leaving campaigns and broadcasters to navigate a state-by-state patchwork.
The industry's calculus
For campaign technologists, the logic is blunt: if the tools work and the other side is using them, sitting out is a competitive liability. The result is an arms race to master a technology that the candidates' own voters say they distrust — campaigns racing to deploy AI in pursuit of an electorate that, in survey after survey, would rather they didn't.
