Pluralism in AI Governance: Toward Sociotechnical Alignment and Normative Coherence

arXiv:2602.15881v1 Announce Type: new
Abstract: This paper examines the challenge of embedding public values into national artificial intelligence (AI) governance frameworks, a task complicated by the sociotechnical nature of contemporary systems. As AI permeates domains such as healthcare, justice, and public administration, legitimacy depends not only on technical correctness but on alignment with societal norms, democratic principles, and human dignity. Traditional paradigms focused on model safety or market efficiency neglect broader institutional contexts. To address this, the study synthesises frameworks including Full-Stack Alignment, Thick Models of Value, Value Sensitive Design, and Public Constitutional AI, alongside comparative analysis of jurisdictions such as the EU, US, China, UK, Brazil, and South Africa (SA). It introduces a layered framework linking values, mechanisms, and strategies, and maps tensions such as fairness versus efficiency, transparency versus security, and privacy versus equity. Findings reveal a pluralism of regulatory philosophies, with SA sovereignty-oriented approach offering a distinctive counterpoint. The study contributes a holistic, value-sensitive model of AI governance, reframing regulation as a proactive mechanism for embedding public values into sociotechnical systems.