Building Collective Intelligence With Artificial Intelligence

I will present learnings from the Global AI Dialogues project, which systematically captures global perspectives on artificial intelligence development and governance. Through structured digital input and deliberative dialogue processes via the Remesh platform, my colleagues and I collected responses from people around the globe, creating a rich dataset exploring values, preferences, and tradeoffs in AI development. Our methodology employs a scenario-based approach across various questions, yielding quantitative and qualitative insights into how different populations evaluate possible AI governance choices. Initial findings reveal interesting patterns, including a robust cross-demographic preference for decentralized AI development and heterogeneous attitudes toward AI delegation that vary by decision context and cultural background. This project serves multiple research and policy objectives: it provides AI labs with empirical data for developing cultural evaluation methods, offers policymakers evidence-based insights into public preferences regarding governance tradeoffs, and enables researchers to analyze global attitudes towards AI. I will discuss implications for incorporating global perspectives into AI governance, methodological limitations, and the role of citizen input in AI alignment. This work contributes to the growing literature on AI governance by providing a replicable methodology for gathering representative global input on critical AI decisions.