Uncovering AI regulation’s public values in the EU

Abstract

Aligning artificial intelligence (AI) with human values is essential to ensure that the technology advances the common good. Governance strategies aimed at fostering human-centric AI range from voluntary ethical guidelines to legally enforceable regulation. Recently, the European Union adopted the AI Act as the world’s first comprehensive AI law. Despite this progress, AI governance frameworks are often criticized for lacking societal grounding, resulting from missing societal participation and measures to assess the impact of AI and its regulation on broader human values and value. Leveraging psychological, non-normative Public Value theory and tools, this paper addresses this gap by refocusing AI governance on the public, its diverse needs and thus a broader notion of value. Quantitative analysis of online survey results (n = 251) reveals that AI regulation provides significant Public Value for citizens along all Public Value dimensions, established in basic human needs. In contrast, qualitative content analysis demonstrates that the AI Act under-emphasizes the political-social and hedonistic-aesthetic dimensions compared to instrumental-utilitarian and moral-ethical dimensions. While the EU’s decision to deviate towards a hard law approach is therefore valuable, actions are recommended to balance out value creation across all basic need dimensions - especially given the identified strong correlation between PV, social trust and intentions to use regulated AI.

Ähnliche Arbeiten