At Stanford's Deliberative Democracy Lab, I'm pursuing independent research on how AI summarization is able to support citizen deliberation. I design experiments that pit different LLM approaches against custom metrics for consensus fidelity and minority view preservation through real-world transcripts, and conduct studies on the interpretation of participants on fairness, legitimacy, and voice. Through advisory from Dr. Alice Siu, I'm ultimately attempting to translate technical design choices into concrete norms for more trustworthy civic technology.