AI Survey Hub and Attitudes Research Exchange

Public opinion research related to artificial intelligence (AI) has dramatically increased over the past several years, tracking the need to better understand attitudes and policy preferences toward a critical emerging technology. However, efforts to produce aggregate insights suffer from challenges related to conceptualization, measurement, data collection, and coordination. In response, we have compiled a comprehensive database of AI-related survey measures that we call AI SHARE, the AI Survey Hub for Attitudes and Research Exchange. To build AI SHARE, we collected and conceptually coded polls, surveys, and survey experiments from academic papers and non-academic sources released. We find that extant research exhibits a geographical bias and fails to examine key subgroups; often conceptualizes AI too generally and overlooks differences in AI knowledge; and lacks reliable and validated measures. We offer concrete recommendations to improve survey research and to enhance coordination between researchers, facilitated by AI SHARE.