AI-generated video of Vermont congressional race tests new state disclosure law - WCAX

An AI-generated video featuring fabricated statements from U.S. Rep. Becca Balint and Republican challenger Mark Coester has circulated on Vermont social media, prompting a debate over political speech and emerging technology. The incident has drawn attention to the effectiveness of recent state legislation designed to regulate the use of artificial intelligence in electoral contexts.

What happened

The video depicts both candidates making statements that they did not actually say. Mark Coester confirmed that he neither produced nor funded the creation of the footage. Rep. Balint described the content as a deepfake, expressing concern regarding the lack of protections for elected officials against unauthorized AI-generated depictions. The clip was shared by Hank Poitras, a Brattleboro-based social media content creator who maintains that the video qualifies as protected satire under the First Amendment.

The video has emerged as a potential test case for a new Vermont law that mandates AI disclosure on political advertisements and videos released within 90 days of an election. Legal experts, including Professor Jared Carter of the Vermont Law and Graduate School, noted that the application of this law hinges on whether the courts categorize the video as political speech or satire, further noting potential constitutional vulnerabilities in the legislation. The Vermont Attorney General’s Office is currently reviewing the video to determine if it falls under the scope of the new disclosure requirements. Rep. Balint, who is also a co-sponsor of the federal No Fakes Act, continues to advocate for increased guardrails to ensure voters can distinguish between authentic communications and synthetic media.