Trust in AI news, AI literacy, and the mediating role of artificial intelligence attitudes: A longitudinal study across diverse societies

Abstract

The rise of generative artificial intelligence (AI) tools, such as ChatGPT, has introduced new ways for people to access and interact with news. Yet, little is known about the factors that shape citizens' trust in AI-generated news. Drawing on a two-wave cross-national panel survey conducted in the United States (n = 1815), Spain (n = 1811), and Chile (n = 1802), this study investigates how AI literacy and attitudes toward AI influence trust in AI-generated news over time. We propose and test a mediation model in which AI literacy affects news trust indirectly through attitudes toward AI. Results show that while AI literacy does not exert a direct influence on trust in AI-generated news, it fosters more positive attitudes toward AI, which in turn enhance trust. This indirect effect is consistent across all three countries, suggesting an attitudinal mechanism linking literacy and trust in AI news trust. Overall, our findings highlight the key role of positive attitudinal evaluations in shaping public trust in communicative AI and underscore that fostering trust requires both improving citizens’ understanding of AI and addressing their broader attitudes toward this technology.