New Jersey’s new and controversial frontier in campaign advertising: artificial intelligence

Josh Gottheimer will never forget the first time he stepped into the boxing ring against Donald Trump. The crowd was roaring, the cameras were flashing, and his toned muscles were rippling; everyone was waiting for him to land the knockout blow against the president and deliver results for New Jersey’s working families.

Did that happen? Of course not. But a new ad Gottheimer is airing in support of his campaign for governor of New Jersey depicts that made-up boxing match and more using generative artificial intelligence (AI) – already the second AI-assisted ad of this year’s gubernatorial race, following a Bill Spadea ad that featured AI voice generation – spurring new questions about the use of AI in political campaigns.

The ad, made by Bill Knapp of the D.C.-based media firm SKDK, features a small disclaimer at the beginning noting its use of AI before launching into images of Gottheimer as a child boxer, Gottheimer standing alongside former Presidents Bill Clinton and Barack Obama, and, finally, Gottheimer in the ring with Trump.

About 80% of the ad is AI-generated, but it also includes some real images of Gottheimer’s days as a Clinton speechwriter. Some, like the Gottheimer-Trump bout, are obviously AI-generated; the origins of others, like Gottheimer standing beside Clinton and Obama outside the Capitol, are not so obvious. It is not clear whether the likenesses of the former presidents are AI-generated or drawn from real photographs and photoshopped together; neither Clinton nor Obama is publicly supporting Gottheimer’s gubernatorial campaign, though the campaign doesn’t necessarily need their permission to legally use their likenesses.

Gottheimer’s campaign did not respond to inquiries about which images were fake and which were real. The congressman also did not answer questions about what made him turn to AI for his advertising in the first place, instead sharing a vague statement saying he would use “every tool in the toolbox” to fight Trump.

“Josh is a born fighter who isn’t afraid to get in the ring to take on Trump when he comes after Jersey families and seniors, and who will fight to lower taxes and costs,” Gottheimer campaign spokesperson Peter Opitz said. “He’s using every tool in the toolbox to fight back against the Trump Administration’s reckless cost-increasing tariff policies and war on Social Security and Medicare.”

It’s not new, of course, for campaigns to feature images of their opponents or of other politicians in their ads; sometimes, the ads use altered or photoshopped images, or utilize actors to portray their opponents in a negative light. Generating other people wholesale, though, is a new frontier only made possible by AI.

The ad, while prompting some incredulity among many New Jersey Democrats, does not seem to run afoul of what minimal regulations there are regarding the use of AI in campaign communications. The Federal Communications Commission proposed a rule last year that would require campaigns to disclose the presence of AI-generated media in advertisements; the rule hasn’t been finalized, and the Gottheimer ad discloses the use of AI anyway.

Legislation is also under discussion in New Jersey to regulate the use of AI in campaign advertisements, but it too simply requires AI-generated content to be clearly noted as such. The driving force behind that legislation, Assembly Majority Leader Lou Greenwald (D-Voorhees), told the New Jersey Globe that he didn’t see anything wrong with Gottheimer’s ad, which discloses its AI use and doesn’t purport to be anything nefarious.

“I think you’re going to know when it’s abusive,” Greenwald said. “I don’t know that I would look at an ad with two individuals in a boxing ring and say, ‘That was the championship belt.’”

Greenwald added that he has a “two-part test” for determining whether an AI ad is out of line: “Is it clearly identified as artificial intelligence? If the artificial intelligence is not being used to distort the position of the individual for personal gain, then it’s proper.”

And as Gottheimer works to break out in the crowded six-way Democratic primary for governor, even the controversy surrounding the ad itself could be a boon. NBC News, which has not been inclined to cover other gubernatorial candidates’ more run-of-the-mill ads, ran a story on Gottheimer’s spot this morning, and there’s been chatter on social media and among top New Jersey Democrats all day about the ethical parameters of AI in politics.

Given the success of other AI-generated content on platforms like YouTube and Facebook, there’s no guarantee all viewers will recognize the ad as featuring fake images, especially if they miss the tiny disclaimer at the beginning. That, naturally, raises questions about where the line is drawn for an AI-assisted ad to be declared as “deceptive.”

But Micah Rasmussen, the director of the Rebovich Institute for New Jersey Politics at Rider University, said that he thought even voters who are aware the ad heavily features AI could still find themselves convinced by its arguments.

“Regardless of whether you think this is satire, regardless of whether you think this happened or not, we want you to see Gottheimer as somebody who could take the fight to Trump,” Rasmussen said, characterizing Gottheimer’s strategy. “And if you see him for that, then it doesn’t matter what else you think of this ad. You got the message.”

Gottheimer isn’t the first New Jersey candidate this cycle to fold AI into their advertising strategy. Spadea, a contender for the GOP gubernatorial nomination, has aired ads featuring an AI recreation of opponent Jack Ciattarelli’s voice reading out an anti-Trump statement that Ciattarelli had written in 2015.

Ciattarelli’s campaign reacted with outrage to the ad, but Spadea’s team argued that since Ciattarelli did in fact write those words, it was fair game to use AI to illustrate that.

“Pay attention to the fact that Jack doesn’t like that a voice is being used to highlight his words – BUT he’s NOT disputing that those words are actually his, because his anti-Trump comments are indisputably his own,” Spadea campaign manager Tom Bonfonti told Politico in November.

And Greenwald said that Spadea’s ad, like Gottheimer’s, passes muster under his own rules: it features a disclaimer stating that AI was used, and it does not use Ciattarelli’s fake voice to distort the former assemblyman’s views (though it may take some liberties with his tone, and it also does not mention that Ciattarelli’s statement was made in 2015). 

The Republican National Committee, too, has waded into AI advertising in the past. An RNC ad from April 2023 showed AI-generated images of the chaos and destruction they argued would have occurred if Joe Biden were re-elected president.

And AI is breaking into the political world in plenty of other ways, too. Former New York Gov. Andrew Cuomo’s mayoral campaign got in some hot water when it was revealed that he had used AI to help write his housing policy plans; Trump himself shared an AI video earlier this year depicting war-torn Gaza as an extravagant resort.

(Some New Jersey candidates have gotten funkier; State Sen. Jon Bramnick, a Republican candidate for governor, shared a fan-created AI rock song advocating for his candidacy.)

But no candidate or campaign has ever put as much money behind an AI-based ad as Gottheimer. NBC News reported Gottheimer’s spot is backed by a “multimillion-dollar” ad buy; since the congressman is receiving matching public funds, and his participation in that program limits his campaign to spending $8.7 million in the primary, that would represent a large share of his total expenditures. (That’s not counting the money spent by a well-funded super PAC supported by Gottheimer’s congressional campaign, though.)

Many past dalliances with AI have drawn mockery or condemnation, and Gottheimer’s ad is no exception thus far. But Rasmussen said that as long as AI-generated ads remain legal, relatively inexpensive, and able to do things that traditional campaign ads cannot, it’s likely to become ever more common.

“It’s not so easy for a campaign to figure out how they want to portray something, and that question becomes a lot easier when the cost and the barrier to visually show something or demonstrate something has been lowered,” Rasmussen said. “So I think we are going to see it more and more.”

The question, Rasmussen added, is: where does one draw the line?

“It is definitely possible that our politics becomes even more detached [from reality],” he said. “We would like to say that there’s a sense of ethical responsibility to realistically portray things, but we all know that lines get blurred on campaigns. When you’re trying to win, that becomes the bottom line: not what’s realistic, not what’s visually clear to voters, but is it going to help you win?”