AI is hurting US politics. Here's how candidates could use it for good
Current US political AI use mirrors outdated one-to-many broadcasting (like old web/TV ads), focused on inundating voters rather than engaging them AI-powered "broad listening" tools can enable genuine one-to-one and many-to-many deliberation at scale, transforming how candidates understand constituent concerns Japan's Team Mirai party has conducted over 16,000 AI interviews with 300,000+ messages, directly citing constituent input in legislative hearings and amending policy platforms accordingl
Analysis
TL;DR
- Current US political AI use mirrors outdated one-to-many broadcasting (like old web/TV ads), focused on inundating voters rather than engaging them
- AI-powered "broad listening" tools can enable genuine one-to-one and many-to-many deliberation at scale, transforming how candidates understand constituent concerns
- Japan's Team Mirai party has conducted over 16,000 AI interviews with 300,000+ messages, directly citing constituent input in legislative hearings and amending policy platforms accordingly
- Scotland's CrownShy nonprofit built the Comhairle platform for digital town halls, while multiple US tools (deliberation.io, Cortico, Remesh, Talk to the City) offer similar open-source capabilities
- A new fund from Higher Ground Labs targeting "AI-Native Campaign Systems" signals growing institutional interest in community-led, bottom-up AI political tools
Why It Matters
This article reframes the AI-in-politics narrative from one of threat (deepfakes, slopaganda) to opportunity, offering practitioners a concrete alternative to manipulative campaign tactics. For AI developers and civic technologists, it identifies a high-impact application domain where open-source tools could reshape democratic engagement. The international examples (Japan, Scotland) provide replicable blueprints that US campaigns could adopt, especially as voter trust in traditional political communication continues to erode.
Technical Details
- Team Mirai (Japan): Built an AI interviewer chatbot that conducts extended conversational dialogues with voters on specific policy issues, integrated with an AI-powered legislative bill exploration portal. The system scales across a wide array of policy topics and feeds synthesized input back to constituents and into formal legislative processes.
- Comhairle (CrownShy, Scotland): Combines AI interviewing with viewpoint synthesis, virtual assembly hosting, and video testimonial sharing to enable mass digital consultations. Funded in part by the Scottish government, designed as open-source civic infrastructure.
- US-based tools: deliberation.io (Stanford-affiliated) facilitates structured dialogues among thousands of participants, piloted by Washington DC; Cortico (MIT-affiliated) surfaces under-heard perspectives from recorded conversations via AI; Talk to the City analyzes large stakeholder input datasets; Remesh generates policy recommendations from AI-synthesized dialogue.
- All highlighted platforms share an open-source philosophy, designed for broad democratic use rather than partisan advantage, continuing the US civic tech tradition from Code for America and the Obama 2008 campaign through the US Digital Service.
Industry Insight
- Campaign technology investors are already pivoting: Higher Ground Labs' new fund targeting AI-native systems and community-led messaging platforms indicates that "listening-first" AI tools may soon become a competitive differentiator rather than a novelty.
- The success of Team Mirai (12 Diet seats) demonstrates that AI-mediated deliberation can translate into real political power, suggesting US candidates who adopt these tools early could capture engaged, underserved voter segments before the space becomes crowded.
- The open-source, non-partisan framing of these tools is a strategic advantage—it builds public trust and avoids the backlash associated with surveillance-capitalism-style microtargeting, making them more sustainable and widely adoptable across the political spectrum.
Disclaimer: The above content is generated by AI and is for reference only.