The growing use of artificial intelligence (AI) in political campaigns has introduced a troubling threat to the democratic process in the U.S., especially with the rise of deepfake technology. This concern was brought into sharp focus in January when, during a Democratic primary, thousands of voters received a robocall that featured an AI-generated impersonation of President Joe Biden, urging them not to vote. The political consultant behind this deceptive tactic now faces severe legal consequences, including millions in fines and the possibility of jail time. He has been charged with 13 felony counts of voter suppression and 13 counts of impersonating a political candidate, a misdemeanor. This incident underscores the growing dangers of AI-generated misinformation in U.S. elections and the urgent need to address it.
In response to this escalating threat, a new Artificial Intelligence Task Force has been established to bring together state and local officials to find solutions. Their aim is to counter malicious AI-generated activity that could potentially undermine the democratic process. The task force, developed by the NewDEAL Forum, is working on crafting policies to mitigate these risks while also leveraging AI’s potential to improve public services.
Arizona Secretary of State Adrian Fontes, an active member of this task force, offered a chilling demonstration of AI’s potential for deception. Though he does not speak German, Fontes created a deepfake video that appears nearly indistinguishable from a real video of him speaking German. The video, intended as a warning, shows how convincingly AI can manipulate reality. “The ability of AI to create lifelike yet completely fake content poses a serious challenge to the integrity of elections,” said Fontes. His warning is clear: as AI evolves, so too must the strategies for safeguarding the electoral process.
Fontes, alongside Michigan Secretary of State Jocelyn Benson and Minnesota Secretary of State Steve Simon, is leading efforts to educate election workers and voters about these AI threats. The coalition of secretaries of state collaborating with the task force is developing best practices for recognizing AI-generated disinformation and responding quickly to prevent its spread.
Michigan has already taken significant legislative steps to protect its elections from AI-generated misinformation. “In Michigan, we’ve enacted legislation to make it a crime for someone to knowingly distribute materially-deceptive deepfakes created with AI that aim to damage the reputation or electoral prospects of a candidate,” Secretary Benson told Democracy Docket, a digital news platform focused on voting rights and elections. This law, passed in November, makes the deliberate use of AI deepfakes for political gain a felony. Additionally, political advertisements created using AI must carry a disclaimer alerting viewers to the ad’s origin. “That disclaimer requirement helps equip citizens with the knowledge of how to be critical consumers,” Benson explained, emphasizing the need for transparency in the age of AI.
Arizona and Michigan, two critical swing states, are taking the threat of AI seriously by organizing tabletop exercises to train election clerks in identifying AI-generated content. These exercises also involve law enforcement and first responders to ensure that any issues arising from AI can be swiftly addressed on or before Election Day.
A NewDEAL Forum poll conducted in Arizona in April revealed that only 41% of respondents were aware of AI’s potential role in elections, highlighting the urgent need for public education on this issue. “Generative AI presents both tremendous opportunities and significant challenges,” said New York State Assemblymember Alex Bores, co-chair of the NewDEAL Forum AI Task Force. Bores, who has a background in computer science, emphasized that their goal is to craft policies that harness AI’s potential for public good while proactively mitigating its risks to democratic institutions.
In March, the task force published a report outlining best practices for election officials at all levels, from secretaries of state to county election workers. These recommendations include public information campaigns to raise awareness about AI threats and protocols for rapid response when disinformation arises. The document also encourages the passage of legislation to strengthen protections against AI manipulation in elections.
As of now, 40 states have introduced legislation to regulate AI use, but only 18 have specific laws addressing election-related AI issues. Michigan’s recent law represents a significant step forward in this regard, setting an example for other states to follow as they grapple with the emerging challenges posed by AI in politics.
