Political Campaigns Ramp Up Use of AI-Generated Ads Ahead of Midterms

Political campaigns are increasingly relying on artificial intelligence (AI) to create highly realistic and often misleading advertisements ahead of the upcoming midterm elections. According to the Wesleyan Media Project, $137 million has been spent on 220 AI-generated ads through October 11th, with most of this spending coming from Republican-aligned groups.

The use of deepfake technology in these ads is raising significant concerns about election integrity and truthfulness in political discourse. In Georgia's Senate race, the Senate Leadership Fund, a super PAC aligned with Republicans, has produced several AI-generated deepfake videos depicting Democratic Senator Jon Ossoff receiving beauty treatments. These ads mock Ossoff’s reported spending on hair and makeup by suggesting that while he "talks pretty," his legislative record is "ugly." Ossoff expressed deep concern over the use of such technology, describing it as a slippery slope toward the destruction of truth in political debate.

In Michigan's Senate race, another super PAC called Front Line Action has used AI to create an ad featuring Democratic candidate Abdul El-Sayed in a fabricated family photo with Hasan Piker and four women wearing niqabs. The ad also includes scenes from a Risk-style board game where cards bear statements like "Death to America," "Do not Celebrate Thanksgiving," and "Celebrate 9/11." Despite the controversial nature of these ads, Front Line Action has defended their use by claiming they are grounded in El-Sayed's public statements and documented reporting.

In New York’s gubernatorial race, Republican Bruce Blakeman’s campaign and allied super PACs have utilized generative AI to produce ads attacking Democratic Governor Kathy Hochul. One such ad features the Tooth Fairy struggling to provide children with money due to financial constraints under Hochul's leadership. Another AI-generated scene shows Hochul at an upscale dinner and outside a police precinct, where she is asked if she wants some crack cocaine. These advertisements do not always disclose their use of AI technology.

Similarly, in Pennsylvania’s gubernatorial race, Republican candidate Stacy Garrity has released an ad featuring an AI-generated version of Democratic Governor Josh Shapiro discussing taxes, abortion rights, and data centers. The ad attributes a fabricated quote to Shapiro, suggesting he fast-tracked data centers after receiving $5 million in political contributions. Shapiro's spokesperson dismissed the ad as "cheap, dishonest AI bullshit."

Currently, only four states have laws specifically prohibiting certain uses of AI-generated deepfakes in political ads, while 27 more require some form of disclosure about their use. However, there is no federal regulation governing the creation and dissemination of such ads.

The increasing prevalence of AI-generated political advertisements highlights a growing disparity between technological advancements and existing election laws. As these tools become more sophisticated and harder to detect, concerns over misinformation and the erosion of public trust in democratic processes are likely to intensify.