Minnesota's Lieutenant Governor and Senate hopeful, Peggy Flanagan, has condemned a recent attack advertisement that features an AI-generated deepfake image of her.
In her remarks, Flanagan expressed her disdain, stating, “Today, you might see a TV ad featuring something that... looks a bit like me. My opponent's super PAC is utilizing an AI deepfake of me to mislead voters. They are unable to win honestly, so they resort to falsehoods. It's appalling. The people of Minnesota deserve better.”
The controversial ad is produced by North Star Dawn, a political action committee backing Congresswoman Angie Craig in her bid for the DFL nomination against Flanagan. It's worth noting that North Star Dawn operates independently from Craig and her campaign. A representative for Craig stated that she does not condone the use of AI in political advertising.
In response to Flanagan's claims, Jerid Kurtz, a spokesperson for the PAC, dismissed the accusations as unfounded. “We will continue to adhere to all federal and state laws while informing Minnesotans that Flanagan's special interest group is funneling millions of corporate dollars into her election campaign,” he remarked in a statement to WCCO.
Minnesota has enacted legislation aimed at regulating the use of AI-generated deepfakes in political contexts to safeguard the electoral process. This law, which was passed during the 2023-2024 session with bipartisan support, stipulates that a deepfake must be so convincingly realistic that a reasonable person would mistake it for official speech or conduct that never actually occurred.
According to the law, deepfakes violate regulations if they:
- Are created without consent from the individual depicted - Are designed to harm a candidate or influence election outcomes - Are released within 90 days prior to a political party’s nominating convention or after absentee voting has commenced for primaries or general elections
Dr. Manjeet Rege, a professor who leads the Center for Applied Artificial Intelligence at the University of St. Thomas, highlighted the increasing difficulty experts face in distinguishing deepfakes from genuine content. “AI-generated material is becoming so lifelike that even experts can struggle to discern what is authentic,” he noted.
Rege also provided tips for identifying deepfakes, suggesting that viewers pay attention to the edges of facial features and hairlines, as well as inconsistencies in background details that may seem off.
