As AI political deepfakes go viral, lawmakers struggle to keep up – 2 states’ laws show limits of what they can do

Thirty-one states now have laws addressing AI-generated election deepfakes, but 28 of them are satisfied if the fabricated video simply carries a warning label saying it was made with AI, according to a June 2026 count by the National Conference of State Legislatures. Only Maryland, Minnesota, and Texas ban certain deceptive deepfakes outright, warning label or not.

Louisiana's law requires disclosure on ads and robocalls using AI-cloned voices of public figures, but exempts federal candidates entirely, meaning a fabricated video of a congressional candidate needs no warning at all. Maryland's law instead requires proof that the deepfake was knowingly spread with intent to influence a vote and caused actual or intended harm, a higher bar that also comes with weaker guarantees the correction reaches voters in time.

By early September 2026, the Wesleyan Media Project had tracked at least 164 AI-made political ads nationwide, accounting for nearly $80 million in spending, with about seven in ten carrying no disclosure whatsoever. Researchers note that warning labels make viewers more skeptical, but there is no clear evidence they change what people share or how they vote.

The full dispatch is available from the source below.