AI is becoming the primary interface for how people form beliefs and participate in democracy, requiring new democratic infrastructure to prevent polarization and protect civic engagement

AI systems are increasingly mediating how people learn what is true, form views on candidates and policies, and act on information. Search is already substantially AI-mediated, and next-generation AI assistants will synthesize information and present it with authority. Personal AI agents will conduct research, draft communications, and inform decisions like how to vote on ballot measures.

Three structural risks emerge: individual AI agents optimized for engagement may produce polarization without explicit political agendas (as seen with social media algorithms); millions of agents interacting simultaneously could generate collective biases even if each agent is well-designed; and a public sphere where everyone has a personalized agent attuned to existing views becomes fragmented private worlds rather than shared deliberation space.

A field evaluation of AI-generated fact checks on X found that people with a variety of political viewpoints deemed AI-written notes more helpful than human-written ones, suggesting AI-assisted fact-checking may achieve cross-partisan credibility. However, the paper is yet to be peer-reviewed.

On the institutional level, policymakers should harness AI's potential to make governance more responsive and legitimate. Several states and localities are already using AI-mediated platforms to conduct democratic deliberation at scale. As agents become increasingly common in public input processes, identity verification for both humans and their agentic proxies must be built in from the start.