The landscape of American political campaigning has entered uncharted territory with the emergence of AI-generated videos, a development that experts fear may become a permanent feature of elections. Recent campaigns in New York City have showcased how this powerful technology is being deployed, sometimes with controversial consequences.
The New York Experiment
While the New York City mayoral election will be remembered for the victory of young democratic socialist Zohran Mamdani, it also marked a significant milestone in political campaigning. Andrew Cuomo, who lost to Mamdani in last week's election, actively shared deepfake videos of his opponent. One particular video led to accusations of racism against the former governor, highlighting the ethical challenges in this new frontier of electioneering.
Alex Bores, a New York state representative pioneering AI regulation laws, observed that this election cycle has broken new ground. "What's really broken through in this election cycle has been the use of generative AI to produce content that goes directly to voters," Bores stated. "Whether that was the Cuomo campaign using ChatGPT to generate its housing plan, or Cuomo and many others making AI-generated video ads for voters, that felt very new in the 2025 cycle."
Controversial Campaign Tactics
The use of AI in the New York race took several concerning turns. Eric Adams, the incumbent mayor who dropped out in September, employed AI to create robocalls featuring him speaking in Mandarin, Urdu and Yiddish. He also produced a dystopian AI video depicting New York as war-torn to attack Mamdani.
Meanwhile, Cuomo faced accusations of racism and Islamophobia after his campaign tweeted a video showing a fictionalised version of Mamdani eating rice with his fingers alongside a Black man shoplifting. The advert also featured another Black man, dressed in purple with a fur coat and silver cane, apparently endorsing sex trafficking. The Cuomo campaign later deleted the video, claiming it had been sent out accidentally.
Bores noted an alarming pattern: "Many of the AI-generated ads in the last election cycle were more likely to veer into what might be perceived as bigoted territory." He suggested this might reflect either algorithms amplifying stereotypes from their training data or the ease of digital manipulation avoiding direct human involvement.
Regulatory Challenges and National Implications
Despite New York state requirements for campaigns to label AI ads, enforcement remains problematic. Some advertisements, including the controversial video Cuomo posted and deleted, lacked proper disclosure. Bores expressed concern that campaigns might willingly accept post-election punishment as a trade-off for potential victory.
The issue has spread to national politics. In July 2024, Elon Musk shared an AI-generated video of Kamala Harris after she became the de facto Democratic presidential nominee. The video depicted Harris claiming she was the "ultimate diversity hire" and admitting she doesn't "know the first thing about running the country."
Robert Weissman, co-president of Public Citizen, emphasised the unique danger of AI-generated content: "Lies have been part of politics since time immemorial. This is different than lies... When someone is shown an apparently authentic version of a person saying something, it is very hard for that person to then contradict it."
With Donald Trump recently sharing an AI video showing him flying a fighter jet and dropping brown fluid on Americans during the No Kings protests in October, and his apparent approval of the medium, prospects for federal regulation appear slim. As midterm elections approach and the 2028 presidential election looms, AI-generated political videos seem destined to remain part of the American political landscape, raising urgent questions about truth, transparency and the future of democratic processes.
