Disinformation Campaigns and the Future of Electoral Integrity in the Age of AI (Policy Brief No. 12)

Harris and Shull's February 2025 policy brief documents how generative AI is reshaping disinformation campaigns targeting electoral integrity, drawing on analysis of the 2024 US presidential election and the broader global electoral cycle. The brief traces the dismantling of platform 'trust and safety' teams as a structural vulnerability — making the argument that platform self-regulatory retreat and AI capability growth are converging to create a significantly more dangerous information environment for elections.

The liability argument is the brief's most contested and most important contribution: Harris and Shull argue that AI companies should be held liable for foreseeable harms produced by their systems in election contexts — a position that, if adopted, would fundamentally change the incentive structure for AI development. The proposals for banning AI impersonation of candidates and mandating watermarking and provenance tools are more immediately achievable and are already informing legislative discussions in multiple jurisdictions. For practitioners working on AI election security and for policymakers designing AI governance frameworks with election integrity dimensions, this brief is current and actionable.