Signal AI and 2026 Elections: Deepfakes and "Liar's Dividend" Threaten Electoral Integrity

Summary

2026 is the second major global election year since generative AI became widely available, and analysts warn the risk has escalated rather than eased. The "liar's dividend," a term coined by legal scholars Bobby Chesney and Danielle Citron, describes how the mere existence of deepfakes lets bad actors dismiss authentic evidence as fabricated. A Brennan Center/CSET "AI and Democracy" essay series, including work by Josh Goldstein and Andrew Lohn, found that heightened awareness of generative AI's power can incentivize politicians to falsely deny the authenticity of real content. The Brennan Center notes that while the worst-case 2024 scenarios of AI-driven disinformation and infrastructure attacks did not fully materialize, AI-generated deepfakes are still eroding trust and polarizing the information environment. Commentary on the 2026 midterms describes the problem as intensifying, with authenticity itself becoming a "partisan preference" that some hope digital watermarks can restore. In response, the Brennan Center has pushed the FEC toward rulemaking on deceptive AI campaign content, and the FEC has issued interpretive guidance that existing fraudulent-misrepresentation rules apply regardless of technology used.