An election is coming. AI is here. We should be careful about how they mix | Opinion

While politicians debate whether voters should have to prove they’re citizens, the rise of AI may prompt voters to make a demand of politicians — prove you’re human.

Artificial intelligence (AI) technology began to play a known role in political campaigns in 2024. This fall, there will be more of it. Campaign regulations and voters’ awareness need to catch up.

Mark Jablonowski, CEO of DSPolitical, a digital advertising company that works with Democratic candidates, said AI is already shaping campaigns.

“I don’t know anyone in politics that isn’t using AI in some way, shape or form, whether it’s to help with scheduling, or to ideate the first draft of some ad scripts, or to helping test which email is going to be most impactful and have the highest donation rate,” he said.

AI lets campaigns do more with less, Jablonowski said. “The one finite quality in a campaign is time,” he said. “If a campaign can run more efficiently, that’s a net positive for the campaign.”

But AI may do more than improve efficiency. Voters may eventually question whether they’re hearing from the actual candidate, or whether the voice on an ad or a recorded call or the content of a stump speech is AI generated. Under the worst scenarios, political opponents could use AI to create videos that fake a candidate’s image and voice.

Jablonowski said AI’s use in campaigns needs guardrails. “AI as a tool can be used positively or negatively,” he said. “We need stronger protections about AI disclosure in advertising and making sure there is still that trust between candidates and the public.”

A glimmer of that confusion appeared in then-Lt. Gov. Mark Robinson’s 2024 gubernatorial campaign. Todd Stiefel, an investor and Democratic donor, created a super PAC – Americans for Prosparody – and launched a nearly $1 million ad campaign using an AI-generated ad targeting Robinson. The ad featured a fake image of Robinson with a version of his voice speaking some of his most controversial statements.

The ad carried a disclaimer: “This satirical ad contains digitally created or altered images and/or audio.” Stiefel defended the manipulation as being based on truth. “The words are his,” he told WRAL-TV. “These are things he’s said and done.”

Morgan Jackson, a political consultant who worked for Robinson’s Democratic opponent — now Gov. Josh Stein — said the AI-generated ad undermined the Stein campaign’s traditional ads targeting Robinson’s statements. “Having somebody fake those videos hurt our credibility,” he said.

Stiefel’s video was based on actual statements, but Morgan said the AI-videos that use false statements or images and are put out by people independent of a candidate could be far more damaging. In those cases, there is no candidate to hold accountable.

“Bad actors who are willing to do or say anything to win an election or steal an election by inserting AI into making candidates appear to be for something, or to have said something, or do something that was not them. That’s where the frontier is. That’s the big concern,” Jackson said.

As AI-generated ads proliferate, it will get harder for the targets of the realistic-looking ads — known as “deep fakes” — to clear the fog of falsehood.

Robyn Caplan, an assistant professor at Duke University’s Sanford School of Public Policy, studies how internet content is moderated. “Deep fakes are extraordinarily cheap to produce now,” she said. ”That means there could be so many produced that that process of debunking is not so effective.”

While fears of AI being misused are legitimate, polling shows the concern is more pronounced among liberals, said Steven Greene, an N.C. State University political science professor. That may have an effect on Democratic campaigns that are reluctant to embrace the new technology, he said.

“In liberal circles, AI is quite frowned upon, but the truth is it’s an amazing tool,” he said. By not fully embracing it, he said, “Democrats may be campaigning with one hand behind their back.”

But it’s a tool that will need heightened voter skepticism, monitoring by traditional media and legal controls. Sen. Amy Klobuchar, D-Minn., and Rep. Yvette Clark, D-N.Y., introduced legislation in 2023 to require clear identification of AI used in political ads. The bill went nowhere.

“This is a genie we can’t put back in the bottle. It’s not going away,” Greene said. “That’s all the more reason we have to be thoughtful about how we use it.”

And alert to how some will abuse it.

Associate opinion editor Ned Barnett can be reached at 919-404-7583, or nbarnett@newsobserver.com

This story was originally published March 29, 2026 at 5:30 AM.