AI Deepfakes Surge in 2026 US Midterm Campaigns — Only 28 States Have Disclosure Laws (2026)

Incident Summary

AI-generated deepfakes surged during the 2026 US midterm campaign cycle, with the National Republican Senatorial Committee (NRSC) releasing an AI deepfake of Texas state representative James Talarico as one of the most prominent documented examples.^[1] Stanford University research documented a broader surge in AI-generated political content across campaigns, finding that AI tools were being used for everything from synthetic candidate imagery to fabricated endorsement videos.^[2] A survey found that 58% of Americans expected AI deepfakes in political campaigns to escalate further. Despite the growing use of AI-generated content in campaigns, only 28 states had enacted disclosure laws requiring AI-generated political content to be labeled, leaving 22 states with no regulatory framework for AI in political advertising.^[3] The regulatory gap means that in nearly half of US states, campaigns can deploy AI-generated content — including deepfakes of opponents — without any legal obligation to disclose the content’s synthetic origin.

Key Facts

• NRSC deepfake: AI-generated deepfake of James Talarico^[1]
• Stanford finding: Surge in AI political content documented^[2]
• Public expectation: 58% expect AI deepfakes to escalate^[3]
• Disclosure laws: Only 28 states have them; 22 states have none^[3]

Threat Patterns Involved

Primary: Disinformation Campaigns — The systematic use of AI deepfakes in midterm campaigns by official party committees represents the institutionalization of AI-generated disinformation in the political process, moving beyond fringe actors to mainstream campaign operations.

Significance

1. Party committee use — The NRSC’s direct use of AI deepfakes demonstrates that AI-generated political content has been adopted by official party campaign committees, not just independent actors or fringe groups
2. 22-state regulatory void — The absence of AI disclosure laws in 22 states creates a regulatory patchwork where the same AI-generated campaign content may be legal in one state and violate disclosure requirements in another
3. 58% public expectation — The majority public expectation that AI deepfakes will escalate suggests growing normalization of synthetic political content, potentially reducing the effectiveness of future disclosure requirements
4. Stanford documentation — Academic documentation of the AI political content surge provides an evidence base for future regulatory action while the campaigns are still active