Advances in artificial intelligence technology are amplifying concerns over how campaigns spread false information, and a partisan deadlock at the Federal Election Commission (FEC) is hindering a progressive-led push to put guardrails in place. 

The FEC is facing a second request — backed by Democrats in the House and Senate — to clarify that its law on fraudulent misrepresentation applies to use of AI.

The new push comes after the commission’s three Republican members defeated an initial petition led by the progressive consumer advocacy organization Public Citizen.

Rep. Adam Schiff (D-Calif.) and Sens. Ben Ray Luján (D-N.M.) and Amy Klobuchar (D-Minn.) sent the FEC’s general counsel a letter last week urging the commission to reconsider its decision, stating that it is “well within” the commission authority to issue AI regulations. 

How AI is changing the 2024 election

“As Members of Congress concerned about the ability of generative AI to significantly disrupt the integrity of our elections, we respectfully request that the FEC reconsider its decision and seek comment on whether the Commission should initiate a full rulemaking on a proposal in the Petition for Rulemaking from Public Citizen. Should you decline this request, please provide a detailed summary and justification as to why you reached that decision,” they wrote. 

The rise of AI in campaigns

Those pushing the FEC to clarify its rules to address AI say it is urgent to take action now in the lead up to the 2024 election, as the technology becomes more advanced and widespread. 

“Artificial intelligence is moving at a rapid clip,” said Lisa Gilbert, executive vice president at Public Citizen. 

Pro-DeSantis group uses AI version of Trump’s voice in new ad 

Gilbert warned that AI can “create images, and audio, and video that is so real that it’s very hard for the viewer to tell that it’s not. And when it comes to elections, that will mean increasingly hard to unpack deepfakes which will be disseminated and could impact the election.”

Some Republican presidential campaigns and super PACs are already using AI as the 2024 presidential primary heats up.

A super PAC backing Florida Gov. Ron DeSantis’s (R) campaign used an AI-generated version of former President Trump’s voice to narrate a post he made on Truth Social. The post attacked Iowa Gov. Kim Reynolds (R), the popular governor of a state crucial to the GOP primary.

Trump had also posted an AI-generated video targeting DeSantis after the Florida governor launched his campaign. And DeSantis’s campaign released an ad that used seemingly AI-produced images of Trump embracing Anthony Fauci, the former director of the National Institute of Allergy and Infectious Diseases. 

Senators grapple with response to AI after first classified briefing

Peter Loge, director of the Project on Ethics in Political Communication at George Washington University, said campaigns have always “cut and pasted” and tried to put their opponent in the “worst possible light,” but the advances in AI are making those images harder to detect and capable of being spread more widely. 

“The question is how do we answer these old questions given the new technologies, and the higher stakes on American elections. And the FEC is part of the answer, Congress is part of the answer, and political consultants and candidates have to be part of the answer,” Loge said. 

Public Citizen sent last week a second petition requesting the FEC clarify that the law against “fraudulent misrepresentation” applies to deceptive AI campaign communications. 

“This is going to have realtime consequences in 2024, and the Federal Election Commission is the agency that takes action to react to this kind of danger,” Gilbert said.

The updated request sought to address concerns raised by GOP commissioners during a June meeting when the request was shot down. Public Citizen sought to clarify why the group believes the commission has the authority to regulate deceptive AI-produced content and specific regulation it wishes to amend. 

Why GOP commissioners are pushing back

During a meeting in June, Republican FEC Commissioners Allen Dickerson, Sean Cooksey and James Trainor III voted to reject Public Citizen’s first petition.

The deadlock between three GOP commissioners and three Democratic commissioners — Chairwoman Dara Lindenbaum, Shana Broussard and Ellen Weintraub — blocked it from going forward.

Dickerson said during the meeting that while he supported an FEC request to expand the commission’s authority, “The only fraud we’re entitled to police is where an agent of one candidate pretends to be the agent of another, or where a person raises funds by fraudulently claiming to be acting on behalf of the campaign with which he or she is unaffiliated.”

“Public Citizen directed its efforts to the wrong component of the government. Instead of coming to us, they should take this up with Congress. I wish it luck,” he added.

Even Lindenbaum, who voted in support of the petition, questioned the panel’s ability to take further action beyond hearing public comment. 

“I also share Commissioner Dickerson’s concerns about whether or not we have any jurisdiction here or the power to do it. I am skeptical that we do, but during the process I hope that we get some … ideas that may help us or help Congress,” Lindenbaum said. 

Can the FEC crack down on AI?

Former FEC Commissioner Ann Ravel, an Obama appointee who served from 2013-17, said the commission has the power to clarify rules for AI.

Although the technology under consideration is newer, considering fraudulent misrepresentation is “not out of the scope of the commission,” she said. 

Meta releases AI tech to rival ChatGPT

Gilbert said Public Citizen is using a “straightforward regulation” to push the FEC to address the use of AI in campaigns.

“What we’re talking about is a new medium that can quickly create that type of misrepresentation and the FEC should interpret their existing reg and address it,” she said.

She said there are “broader” regulations she would want to see, such as requiring watermarks or clarifications to identify AI, that the FEC does not yet have the authority to put in place.

Ravel said it would have a “significant” impact if the FEC was able to take action on AI — or even attempt to do so — as a tool to influence the companies that are producing it. 

“I know there’s a lot of people that think there’s good from it, but a lot of people that are really scared as well. And so having them come out with this, because of the concerns about the electoral process, I think would be a meaningful act on their part, and should be easier than getting a majority of Congress to vote for it,” Ravel said.

Although, she said, the FEC’s deadlock and apparent wait for express authority from Congress is “typical” for the agency given the 3-3 partisan split and need for four votes to take action. 

Loge also cautioned that as regulators and lawmakers weigh AI rules, they should be wary of viewing the debate in “technological extremes.” 

“This is neither the end of the world, nor is it going to lead to a utopia full of puppies and rainbows. Emerging technology tends to be a bit more complicated, and tends not to have the extreme benefits or downfalls that people hope for or fear,” he said.