US Midterms Face AI Deepfakes as 300,000 Japanese Chatbot Messages Show Better Campaigning Path
Updated
Updated · The Guardian · Aug 31
US Midterms Face AI Deepfakes as 300,000 Japanese Chatbot Messages Show Better Campaigning Path
1 articles · Updated · The Guardian · Aug 31
Summary
US midterm politics is absorbing AI mainly as a broadcast weapon—deepfakes, misinformation and White House “slopaganda”—while candidates largely ignore tools that could improve voter listening and policy feedback.
One promising model comes from Japan, where Team Mirai says voters have sent more than 300,000 messages through 16,000 AI interviews that feed into bill discussions, public syntheses and even platform changes.
That approach contrasts with standard US campaign tactics by using AI for one-to-one and many-to-many deliberation rather than saturating voters with ads and talking points.
Open-source systems from Scotland’s CrownShy and US-linked projects including deliberation.io, Cortico, Talk to the City and Remesh show the tools already exist for large-scale constituent consultation.
The broader argument is that AI in the 2026 midterms need not split neatly into booster-versus-skeptic camps if candidates separate the technology from big-tech interests and deploy it for more responsive campaigning.