State Responses to AI-Driven Election Misinformation: Governing Election Integrity in the United States

• In Event: Politics and Policymaking in State-Level AI Governance

Abstract

Generative artificial intelligence has created a new governance challenge for U.S. election administration by lowering the cost of producing deceptive synthetic media, impersonation content, and misleading election information. These dynamics heighten risks to voter trust, administrative integrity, and information environments during campaigns and early voting periods. Because election administration in the United States is highly decentralized, state governments have been playing the key role of policy innovation and institutional response. However, the fragmented election governance structures also leaves us less knowledgeable about how states differ in building governance capacity to address AI-related election misinformation threats.
The study introduces an original fifty-state qualitative dataset of state election responses compiled from statutes, administrative guidance, secretary of state materials, attorney general resources, official press releases, policy documents, and verified secondary reporting. Each state is coded across AI-relevant dimensions of election governance, including disclosure and prohibition rules for synthetic media, enforcement and remedial mechanisms, AI-focused preparedness and coordination capacity, public guidance and reporting systems, and technical authenticity or provenance tools. These qualitative data are paired with a four-level maturity framework to construct a comparative AI Election Governance Index and enable a systematic cross-state comparisons over time.
Preliminary evidence shows substantial interstate variation. Some states have adopted more developed governance frameworks that combine legal regulation, administrative preparedness, and technical verification tools, while many others rely primarily on general election-security or public-information measures that are not tailored to AI-specific threats. Advanced technical authenticity and provenance systems remain limited to a relatively small set of cases.
The paper contributes to research on state policy variation, public-sector innovation, and digital governance by showing how decentralized governments are adapting election administration to the risks posed by generative AI. It also offers a practical framework that election officials and policymakers can use to benchmark and strengthen state-level AI governance capacity.

Author

• Jeongmin Oh, Hanyang University
  Presenting Author