Opportunities and Risks of LLMs for Scalable Deliberation with Polis

Research exploring how LLMs (specifically Claude) can meaningfully enhance Polis conversations, demonstrating that summarization capabilities enable categorically new methods with immense promise to empower the public in collective meaning-making exercises. While highlighting substantial benefits for moderating, facilitating, and summarizing discussions, the paper acknowledges context limitations significantly impact quality and proposes principles and mitigation strategies. The research outlines frameworks for addressing pitfalls and identifies open questions for AI-augmented deliberative platforms.