Artificial Intelligence (AI) has rapidly moved from expert arenas into everyday life, acquiring strong symbolic and cultural relevance within contemporary social conversations around science. Despite increasing familiarity with AI technologies, public attitudes remain ambivalent and polarized, challenging deficit-based assumptions that link scientific knowledge to positive technological orientations. Adopting an interpretative Public Understanding of Science framework, this study examines how scientific literacy, trust in science, and exposure to AI-related content interact in shaping public attitudes toward AI. The analysis draws on data from the 2025 Observa Science in Society Monitor, a representative survey conducted in the Emilia-Romagna region (n = 502). Latent Class Analysis identifies four distinct attitudinal profiles—techno-ambivalent, techno-optimist, techno-skeptic, and techno-phobic—which are subsequently analyzed using multinomial logistic regression. Results show that trust in science is the most robust predictor of techno-optimistic attitudes, outweighing the effects of scientific literacy and media exposure. Moreover, exposure to AI-related content has conditional effects: among individuals with low trust in science, higher exposure increases the likelihood of techno-phobic orientations, while this association disappears among high-trust respondents. Overall, the findings highlight the central role of trust in science in shaping differentiated public responses to emerging technologies beyond deficit-based explanations.
Between Trust and Anxiety: Citizen Attitudes to AI in Emilia-Romagna
Andrea Rubin
2026
Abstract
Artificial Intelligence (AI) has rapidly moved from expert arenas into everyday life, acquiring strong symbolic and cultural relevance within contemporary social conversations around science. Despite increasing familiarity with AI technologies, public attitudes remain ambivalent and polarized, challenging deficit-based assumptions that link scientific knowledge to positive technological orientations. Adopting an interpretative Public Understanding of Science framework, this study examines how scientific literacy, trust in science, and exposure to AI-related content interact in shaping public attitudes toward AI. The analysis draws on data from the 2025 Observa Science in Society Monitor, a representative survey conducted in the Emilia-Romagna region (n = 502). Latent Class Analysis identifies four distinct attitudinal profiles—techno-ambivalent, techno-optimist, techno-skeptic, and techno-phobic—which are subsequently analyzed using multinomial logistic regression. Results show that trust in science is the most robust predictor of techno-optimistic attitudes, outweighing the effects of scientific literacy and media exposure. Moreover, exposure to AI-related content has conditional effects: among individuals with low trust in science, higher exposure increases the likelihood of techno-phobic orientations, while this association disappears among high-trust respondents. Overall, the findings highlight the central role of trust in science in shaping differentiated public responses to emerging technologies beyond deficit-based explanations.I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
