Why Banning AI Tools Won't Stop Election Deepfakes, Experts Say

Election officials warn that cheap, convincing AI-generated media is already being used to spread false voting claims, and new deep-fake bans may miss the mark.

The upcoming November midterms will be the first national contests where inexpensive, high-quality generative AI is widely accessible, prompting concerns about fabricated election evidence. Los Angeles County officials confirmed that AI-manipulated video of ballot counting was deployed to stir fraud narratives during the June primary. Although thirty-one states have enacted deep-fake bans, many are expected to falter because statutes either target specific algorithms that can be swapped out or overreach to include ordinary editing tools.

Federal and state courts have already invalidated broad prohibitions as unconstitutional restrictions on political speech, while narrower disclosure requirements have held up. Authorities have successfully prosecuted AI-driven deception using existing voter-suppression and impersonation laws, exemplified by a recent AI-cloned Joe Biden robocall in New Hampshire that led to an indictment and a $6 million FCC fine. Experts recommend focusing on the deceptive act rather than the tool, applying consistent political-ad disclosure rules, and encouraging AI firms to enforce anti-impersonation policies. The core message is that the legal response should target lies, regardless of whether a human or a machine creates them.