Many Americans say they’ve been confused or even fooled by political ads using artificial intelligence this campaign season, and want more rules governing how the technology is used, according to a new Scripps News/Talker Research poll.
More than four-in-ten registered voters say they’ve seen something they thought was real in a political ad, only to later find it was AI-generated. Another 30% said they saw something they weren’t sure was real or AI.
Some of the most talked about ads of the campaign have featured AI-generated candidates like Sen. Jon Ossoff in a beauty salon, James Talarico reciting old quotes out of context, and Sherrod Brown overstaying his welcome at a birthday party.
More than half of survey respondents (56%) said they have seen Republicans using AI ads, compared with 49% saying they have seen ads for Democrats.
IN CASE YOU MISSED IT | 'I had the flu': Talarico returns to campaign trail, dismisses online rumors
And more than three-quarters of voters say there should be rules around how the technology is used in politics.
In the absence of federal regulation, states have passed their own laws aimed at curbing the influence of AI or deepfakes in campaign advertising. Some states require disclosures or bar deepfake videos from being published in the weeks leading up to Election Day. But many of the regulations fail to keep up with the evolving use of what can be a cheap and effective new tool.
An analysis from the Wesleyan Media Project from early September found AI tools had been used in at least 164 ads, 69% of which did not disclose AI was used. Eighty percent were from Republicans or Republican-leaning groups.
Still, a similar amount of voters in both parties say there should be rules around how the technology is used, according to the Scripps News/Talker Research poll, including 77% of Democrats and 79% of Republicans.
Concerns about how AI could impact areas outside politics also cut across party lines. Large swaths of the country worry about the impact AI could have on their jobs (61%), government institutions (83%), financial institutions (81%), and their physical safety (57%).
Nearly 80% say they see content they believe was generated by AI at least once per day.
This random double-opt-in online survey of 1,000 U.S. registered voters was commissioned by Scripps News and conducted by market research company Talker Research, in accordance with the Market Research Society’s code of conduct. Data was collected from Sept. 30-Oct. 1, 2026. The margin of error is +/- 2.2 points with 95% confidence
