Risk, Trust, and Transnational Media Exposure: Explaining Support for the EU AI Act among Foreign-Born Residents in Sweden

Sammanfattning

This study investigates how foreign-born residents in Sweden perceive the European Union Artificial Intelligence Act (EU AI Act) and express individual-level attitudinal support for it within transnational information environments. To address this gap, this thesis develops an analytical framework integrating transnational media repertoires, source-based media-repertoire indicators, value-frame identification, institutional trust, and privacy calculus. Empirically, the study uses a cross-sectional online survey of 78 foreign-born residents living in Sweden. The analysis addresses three research questions. Firstly, to what extent transnational media-repertoire indicators are associated with value-frame identification; secondly, to what extent media-repertoire indicators and institutional trust explain support for the EU AI Act; and thirdly, to what extent privacy-related risk-benefit evaluations are associated with support after accounting for media-repertoire indicators and institutional trust. The results provide limited support for the direct role of media-repertoire indicators. These indicators were not significantly associated with either privacy-and-rights-oriented or innovation-and-competitiveness-oriented value frames. They also did not significantly explain support for the EU AI Act in the regression models. By contrast, privacy calculus substantially improved the explanatory power of the final model. Respondents who evaluated AI more in terms of risks than benefits expressed stronger support for the Act. Institutional trust was also positively associated with regulatory support. The thesis therefore suggests that support for AI regulation in this sample is better understood through evaluative mechanisms than through broad differences in media exposure alone. Transnational media repertoires, as approximated by media-repertoire indicators, remain relevant as contextual conditions for information exposure, but support for the EU AI Act is more directly associated with how respondents evaluate AI-related risks, benefits, and institutional capacity.