AI Deepfakes Proliferate in US Election Campaigns

Artificial intelligence-generated deepfakes have become a prominent feature in contemporary US election campaigns, moving beyond theoretical warnings to practical, often juvenile, applications in political discourse. This trend is particularly evident in recent primary elections, where AI-driven content is being used to influence voter perception and attack candidates. In a notable instance during a Michigan Senate race, an attack video supporting Republican candidate Mike Rogers depicted Democratic nominee Abdul El Sayed alongside Alexandria Ocasio-Cortez and Bernie Sanders, employing a Trojan horse metaphor. This video garnered at least two million views on the social media platform X. The use of AI in political advertising extends to other races. In the Massachusetts gubernatorial contest, the campaign of Republican candidate Brian Shortsleeve released AI-generated videos critical of incumbent Governor Maura Healey. One such video, themed for Halloween, portrayed Healey as a vampire in a "house of horrors," while another mimicked her voice in a satirical "radio ad." These tactics are part of a wider pattern. The National Republican Senatorial Committee (NRSC) utilized an AI-generated version of Texas Senate candidate James Talarico to recite his past tweets, some of which were considered polarizing. Furthermore, following New York City Mayor Zohran Mamdani's electoral victory, anonymous social media accounts deployed AI to create synthetic local voters who voiced criticism against him. Efforts to regulate the use of deepfakes in elections are underway at the state level. The National Conference of State Legislatures tracks numerous state statutes aimed at restricting or prohibiting the creation and dissemination of AI-generated content in political contexts. These laws address various scenarios, including intentionally deceptive AI and the creation of digital replicas without an individual's consent. Specific state regulations include disclosure requirements for AI used in political advertisements during election periods, such as in Arizona. Other states, like Minnesota, have enacted prohibitions against nonconsensual deepfakes intended to damage candidates' reputations. Despite these state-level initiatives, a comprehensive federal law specifically governing deepfakes in elections remains absent, leaving a regulatory gap in this evolving landscape. The proliferation of these AI tools raises concerns about the integrity of electoral processes and the potential for widespread misinformation.