A video from the Ohio Republican Party showed former Vice President Kamala Harris playing a game show with Dr. Amy Acton, the Democratic nominee for governor.
Spoiler alert: That never happened. It was AI-generated content intended to mock Acton’s past comments. And it won’t be the last time Ohioans see deepfakes on the campaign trail.
Experts say artificial intelligence is exploding in the political sphere as technology becomes more sophisticated. Generative AI in particular − which allows users to create text, photos, video and audio − is opening doors for campaigns and other groups that want to find efficiencies and save money.
But it’s also sparked debate over the role of AI in elections and whether it will help or hurt democracy.
“Things like playing around with the boundaries of misleading content, things like borderline falsely depicting your opponents or trying to emotionally appeal to or manipulate people while technically following regulations, I don’t think these are things that political campaigns should be lauded for,” said Daniel Schiff, co-director of the Governance and Responsible AI Lab at Purdue University.
“It’s a dangerous game they’re playing.”
How political campaigns deploy AI
AI in politics is not a new phenomenon.
Schiff’s organization has been tracking examples since 2016. In 2024, for example, Donald Trump Jr. shared an AI image of his father riding a large cat, gun in hand, with the caption, “Save our pets!!!!!” The post was a reference to false, viral claims about Springfield, Ohio’s Haitian immigrant population.
Still, AI has accelerated with the advent of ChatGPT and other tools that make it easy and accessible for anyone to create content. Many of these photos and videos are created by everyday people, but it’s becoming more common for political actors to give it a try.
“Our party leadership gave us room to innovate, and we ran with it,” said Nick Johnson, content and communications director for the Ohio Republican Party. “I’m proud to lead that effort every single day. We are confidently outpacing Ohio Democrats 6 to 1 on content production, and every ounce of that output is aimed at one thing: winning in November.”
Last month, a PAC supporting Sen. Jon Husted aired an ad critical of his opponent, former Sen. Sherrod Brown, that appeared to be made with AI. It also cropped up in ads for the contentious GOP primary between northern Kentucky Congressman Thomas Massie and Ed Gallrein.
Schiff said another common tactic is using voice manipulation to depict politicians, which can be persuasive to voters − real or not.
“This content influences people,” Schiff said. “It draws on their emotions. It can affect their views.”
Campaigns aren’t just using AI to make ads and memes. Grassroots Analytics helps progressive political groups aggregate news and create press releases, mailers and fundraising messages, among other written content. It doesn’t offer image generation.
CEO Meghan McAnespie said these tools save campaigns time and allow them to tailor messaging to specific voters based on demographics, geography or issues. She believes AI can open doors for candidates who previously didn’t have the money or resources to run for office.
At the same time, McAnespie worries that progressives will cede a technology advantage to conservatives because of their concerns about AI, such as its impact on the environment.
“It’s always dangerous when there’s an imbalance in tech and tools available or all of the resources being leveraged there,” McAnespie said.
Ohio bill would require AI campaign disclosures
As campaigns experiment with advancing technology, Ohio lawmakers are debating how to regulate it.
House Bill 185 would require disclaimers in political ads that use deepfakes, including information about the entity behind the content. Rep. Adam Mathews, R-Lebanon, said AI will inevitably become part of campaigns, but voters deserve to know what’s real and what’s not.
“People need to know if those videos or pictures actually happened so they can make informed decisions on the reality around them,” said Mathews, who co-sponsored the legislation.
Twenty-seven other states require political groups to disclose AI, according to the National Conference of State Legislatures, while Minnesota and Texas ban deepfakes when an election gets closer. Laws generally don’t go further because of free speech concerns, although there’s some debate about whether AI is liable for defamation.
But campaign regulations can’t stop random social media users from sharing AI content that affects political discourse.
“It’s tough to say there should be some teenager held legally liable because they made a video of a candidate doing something silly,” said Ryan Kennedy, a political science professor and data analytics leader at Ohio State University. “That kind of stuff is just going to be more and more common.”
State government reporter Haley BeMiller can be reached at hbemiller@usatodayco.com or @haleybemiller on X.
What’s your view?
Do you think AI should be used in political ads? Tell us what you think
This article originally appeared on The Columbus Dispatch: Ohio voters will see more AI political content this election cycle
Reporting by Haley BeMiller, Columbus Dispatch / The Columbus Dispatch
USA TODAY Network via Reuters Connect
