AI-Generated Fakes Seen as Voter Suppression Tool
AI-driven disinformation and deepfakes are increasingly being used as tools for voter suppression and intimidation ahead of the US 2026 midterm elections. An analysis warns that synthetic media has evolved beyond simple misinformation to become a weapon that could disproportionately affect marginalized communities and swing voters.
- A prominent real-world example of AI-driven voter suppression involved AI-generated robocalls that mimicked President Joe Biden's voice ahead of the 2024 New Hampshire primary, falsely urging Democrats to "save your vote" for the November election. The political consultant behind the scheme was later fined $6 million by the FCC. - The cost and speed of creating this type of disinformation are remarkably low; the fake audio for the New Hampshire robocall reportedly took less than 20 minutes and cost only $1 to produce. - Federal law does not specifically regulate AI in political campaigns, but the Federal Communications Commission (FCC) has affirmed that existing rules against "artificial or prerecorded voice" robocalls apply to AI-cloned voices. This, however, does not stop political robocalls to landlines, which are still permitted without prior consent. - There is a growing but inconsistent patchwork of state-level legislation. As of early 2026, 26 states have enacted laws to regulate political deepfakes, with most requiring disclosures and two states—Minnesota and Texas—prohibiting them entirely within a certain window before an election. - Legal challenges to these regulations are emerging, citing First Amendment rights. A 2024 California law that prohibited political deepfakes was struck down by a federal court in August 2025 as being overly vague and not regulating actual harm. - Suppression tactics are increasingly aimed at non-English speaking and minority communities, where AI can quickly generate convincing, in-language disinformation that is difficult to track and counteract. - Beyond impersonating candidates, AI can be used to create and rapidly spread false information about polling locations and times, or to generate chatbots that mislead voters on how to cast their ballots. - A key challenge is the "liar's dividend," a phenomenon where the widespread knowledge of deepfakes allows political figures to dismiss authentic, damaging audio or video evidence as being fake, further eroding public trust.
