This research reveals how generative AI can enable large-scale democratic participation while maintaining fairness despite inherent LLM limitations.

• Multiple LLM-generated votes can produce fair collective outcomes even when individual AI responses show bias
• Voting aggregation mechanisms effectively neutralize inconsistencies in AI-generated preferences
• The approach demonstrates resilience against manipulation and ensures democratic representation at scale

Security Implications: This work provides critical safeguards for deploying AI in democratic processes, establishing that collective AI voting can resist manipulation while protecting democratic integrity—even when individual AI agents exhibit biases.

Generative AI Voting: Fair Collective Choice is Resilient to LLM Biases and Inconsistencies