Trust in Artificial Intelligence and AI-Assisted Social and Civic Decision-Making Among Arab Youth: Evidence from Egypt, Kuwait, and the UAE

Abstract

Introduction: Trust in artificial intelligence (AI) increasingly shapes how young people evaluate information and public issues, yet evidence from Arab contexts remains limited. This study examined the socio-technical correlates of trust in AI and its association with AI-assisted social and civic decision-making among digitally engaged youth in Egypt, Kuwait, and the United Arab Emirates. Methods: A cross-sectional online survey was conducted with 1,023 respondents aged 18–35 recruited through purposive, non-probability sampling. Confirmatory factor analysis (CFA), structural equation modeling (SEM), multi-group CFA, and competing full-versus partial-mediation models evaluated measurement and structural relationships. Results: The partially mediated model fit significantly better than the fully mediated model, Δχ²(5) = 55.66, p < .001. Institutional confidence (β = .303), AI literacy (β = .281), perceived usefulness (β = .242), and social influence (β = .107) were positively associated with trust in AI, and perceived risk was negatively associated (β = −.106; all p < .001). Trust was positively associated with AI-assisted decision-making (β = .412, p < .001). Significant indirect associations through trust were observed for all five antecedents, while the direct perceived-risk path to AI-assisted decision-making was nonsignificant. Multi-group CFA supported metric and practical scalar measurement invariance across the three countries. Discussion: The findings support a socio-technical account of AI trust in which technological, cognitive, institutional, and social conditions jointly relate to AI-assisted decision-making, with trust functioning as a central but not exclusive pathway. Given the cross-sectional, purposive design, findings are interpreted as associations rather than causal or nationally representative effects.