Societal promises and perils of AI: Sociotropic perceptions, anxiety, and structural filters

Research on AI anxiety has often assumed that perceptions of AI's societal risks and benefits translate uniformly into emotional responses, regardless of social position. Drawing on survey data from 2023, we combine social stratification theory with the sociology of risk to conceptualize AI anxiety as a sociotropic concern filtered by structural position. We examine how individuals simultaneously perceive societal benefits and risks of AI, and how AI familiarity, political ideology, and subjective economic wellbeing shape the relationship between these perceptions and anxiety. The results show that perceived benefits and risks operate as distinct and independent predictors of anxiety, with benefits exerting the stronger influence. Structural filters moderate this relationship. Greater AI familiarity amplifies both effects, a pattern that challenges knowledge-deficit models and that we interpret as informed differentiation. Political ideology selectively filters benefit perceptions, while economic vulnerability limits the extent to which individuals derive reassurance from them. By contrast, the effect of perceived societal risks on anxiety does not vary across ideological and economic lines. These findings reframe AI anxiety as a stratified sociotropic response and suggest that information-based governance strategies may prove insufficient without differentiated approaches that address structural vulnerabilities and distributional consequences.