The First AI Election? How AI Is Already Totally Reshaping Politics

2026 is the first American election where AI is shaping every link in the chain of political power — from what campaigns put in front of voters, to how voters filter what they're told, to who tech companies are trying to elect.Matt Robison sits down with Katie Harbath — 25-year veteran of the tech-politics intersection, former Facebook elections lead, and author of the upcoming book Disrupting Politics: A Front Row Seat at the Collision of Technology and Democracy (Fall 2026) — to walk through her original survey data and what she's seeing on the ground right now.We cover:Her first-of-its-kind survey of campaign professionals: 87% use AI daily, and most campaigns still have no policy governing itWhy disclosure is more complicated than "just label it" — and the research showing labels can backfireThe voter side: only 9% say they use AI for political info, but that number is misleading once you count AI answers baked into a Google searchWhy Democratic campaign staff lean Claude and Republicans lean ChatGPT and GrokHard lessons from 2016 — and the new "flood the zone" risk Katie's most worried about heading into 2026 and 2028The most underrated vulnerability in American elections: the fog-of-war period between election day and certificationWhy state legislatures — not Congress — may be the most exposed to AI industry money and influence right nowThe forgotten story of Congress eliminating its own in-house science expertise in 1994, and what it has to do with today's AI policy gridlockKatie's own verdict on whether this is the first AI election — and possibly the last one where AI isn't the dominant issue📬 Subscribe to Worth Knowing: https://worthknowing.substack.comKatie Harbath writes Anchor Change, covering the intersection of tech, politics, and democracy: https://anchorchange.substack.com⏱️ TIMESTAMPS00:00 Stream picks up mid-conversation (technical difficulties at the top — sorry, folks)00:41 Recap: the four links in the AI-elections chain02:56 "Panic responsibly" — Katie's framework for AI fear vs. AI promise05:00 Inside Katie's first-of-its-kind campaign AI survey07:00 Is campaign AI use actually bad? Logos vs. deepfakes10:07 How voters are using AI to filter political information11:47 The Forum AI bias research and the "information gap" problem13:11 Lessons from 2016 at Facebook — and the new "flood the zone" risk17:34 Do Anthropic and OpenAI run election war-room scenarios?20:32 The most dangerous window: election day to certification22:53 The Detroit Edison hypothetical and shaping post-election perception25:00 Audience Q&A: why doesn't Facebook stop AI deepfakes?28:56 Geoffrey Hinton's fear: can we still tell reality from unreality?33:14 The optimism case — news audits, agency, and the return to in-person36:14 Audience Q&A: is Kara Swisher's tech reporting reliable?37:16 AI money in elections — and why state legislatures are most exposed41:31 The forgotten story of Congress eliminating its own science office44:47 The politicization of AI policy in Washington46:56 Verdict: is this the first AI election — and the last non-AI one?49:29 Wrap-up and what's next🔑 KEYWORDS: AI and politics, AI in elections, first AI election, Katie Harbath, campaign AI survey, AI deepfakes, Claude vs ChatGPT politics, AI money in elections, state legislature AI influence, election disinformation, Disrupting Politics book, Worth Knowing podcast#AIandPolitics #KatieHarbath #2026Midterms #ArtificialIntelligence #WorthKnowing #DisruptingPolitics #TechPolicy #Deepfakes #ElectionSecurity #CampaignTech