AI & Elections / Synthetic Political Media

The strongest good-faith objection, already partly conceded in this issue's strategy layer: the narrow scope, precisely because it's designed to survive First Amendment challenge, leaves a large and arguably more consequential category of election-deepfake harm (fabricated candidate statements, fake endorsements) completely unaddressed at the federal level. A sophisticated bad actor could deliberately stay just outside the voting-mechanics carve-out and cause election harm this law does nothing about. That's true, and it's exactly why this is the right scope rather than a lesser one: a broader federal law covering candidate- statement and endorsement deepfakes would almost certainly draw the same First Amendment challenge that has already struck down every broader state version, protecting no one while the case is litigated. A narrower law that survives and deters voting- mechanics disinformation delivers more protection than a broader one that reads better on paper and gets enjoined before the election it was supposed to protect.

Voters exposed to non-voting-mechanics deepfakes (fabricated candidate statements, fake endorsements) get no federal-law protection under this issue's scope. Candidates targeted by such deepfakes bear reputational harm with no federal remedy this issue provides. States attempting broader protections bear ongoing litigation costs and legal uncertainty that a narrow federal floor doesn't erase. That doesn't mean this position asks them to stop trying broader state-level protection; it only argues the federal law itself should be the version built to survive, which is the more useful contribution a federal statute can make to a fight states are already going to keep having regardless.