Illustration: Shoshana Gordon/Axios
AI-generated political ads could explode in upcoming elections as generative AI becomes more popular, adding a serious complication to online advertising that already struggles with misinformation.
Why it matters: No federal rules are on the books about the use of AI in political advertising — though that may be about to change.
Driving the news: The Federal Election Commission announced Tuesday it will hold a meeting Thursday to consider whether to initiate a full rulemaking on political deepfakes.
- In May, Public Citizen petitioned the FEC to clarify that its existing law against "fraudulent misrepresentation" applies to deceptive campaign ads using AI.
The big picture: Right now, there's a vacuum. Major tech platforms that sell and host political ads have yet to develop definitive new policies around it.
- "This is an issue that's going to continue to snowball," said Stephen Spaulding, VP of policy and external affairs at Common Cause. "There are tools the FEC could employ, but no question, there also has to be a comprehensive legislative response."
- Katie Sanders, managing editor of PolitiFact, told Axios that it's "alarming to see more candidates embracing this technology" — especially video and audio manipulation, since it's "just kind of created out of whole cloth, where you don't have something original to reference."
- Emma Steiner, disinformation analyst at Common Cause, told Axios that the group's biggest concern about generative AI is that platforms "are not willing to act on evolving disinformation narratives."
- "So while [companies] can attempt to drop new policies ... I'm not sure they will make an actual serious attempt to counter the issue."
Driving the news: A new ad from Republican presidential candidate Ron DeSantis, governor of Florida, features AI-generated, fake images of former President Trump kissing and hugging Anthony Fauci.
- A disclaimer on the video, written by users and shared on Twitter, says the video contains AI-generated elements. But critics say those types of disclaimers are insufficient.
- An AI-generated Republican National Committee ad from April portrays a dystopian future if President Biden is re-elected, Axios previously reported.
- Campaign managers and digital strategists are already thinking about how AI will affect the 2024 election, per Axios' Sophia Cai.
Tech takes: "Since 2018, we have provided industry-leading transparency for ads about social issues, elections or politics," Kevin McAlister, a Meta spokesperson, told Axios.
- "We remain committed to that goal, and we are working with experts to determine what, if any, changes are needed to address the use of generative AI."
- Snapchat spokesperson Pete Boogaard told Axios: "We use human review to fact-check political and advocacy ads and have strict policies that help mitigate the risk of deceptive or false information in ads."
- "We regularly evaluate our policies to make sure our protections keep pace as technologies evolve, including AI, and will continue to do so."
- Michael Aciman, a Google spokesperson, told Axios: "All ads on our platform, including election ads, must comply with our policies regardless of how they’re created. These policies already prohibit the use of manipulated media to mislead people, like deep fakes or doctored content, as well as ads that promote demonstrably false claims about elections."
- "We’re closely monitoring the use of generative AI in ads to determine how best to improve transparency for users."
State of play: Sens. Amy Klobuchar, Cory Booker and Michael Bennet raised the alarm in May, introducing a Senate bill (with a House version sponsored by Yvette Clarke) that would require a disclaimer on political ads that use images or video generated by AI.
- "Artificial intelligence has the potential to completely upend campaigns and elections unless there are rules of the road in place," Klobuchar said in a statement to Axios. "I’m pushing to immediately pass stronger laws regarding AI-manipulated content to counter the spread of election-related disinformation."
- The Democratic National Committee declined to comment. The Republican National Committee didn't respond to a request for comment.
Flashback: In 2019, then-candidate Joe Biden pledged to not use deepfakes, synthetic social media accounts or bot networks in his campaign.
- "This is a place where there could be some bipartisan momentum around tech accountability because it's something that both major parties are really going to potentially have to grapple with," said Spaulding at Common Cause.
