Abstract The rapid proliferation of artificial intelligence (AI) technologies across corporate value chains has exposed fundamental gaps in traditional approaches to ensuring respect for human rights. This article analyzes the transition from fragmented compliance mechanisms to unified governance models that seek to embed human rights standards within AI deployment, focusing in particular on the European Union’s regulatory framework and its global implications. A central focus is placed on the EU’s AI Act, which introduces the Fundamental Rights Impact Assessment (FRIA) as the first mandatory evaluation mechanism for selected high-risk AI systems. We demonstrate that FRIA marks a significant regulatory innovation. Alongside FRIA, we examine the Council of Europe’s voluntary HUDERIA methodology, which while conceptually aligned with FRIA, enables organizations to undertake comprehensive rights assessment across their entire AI portfolio. We argue that the convergence of mandatory (FRIA) and voluntary (HUDERIA) frameworks strengthens corporate responsibility for human rights in AI, extending ethical obligations where legal norms may not reach. However, we also highlight the real-world challenges to universalizing these standards – growing regulatory divergence between the EU’s rights-driven approach and the US innovation-focused “America’s AI Action Plan.” The resulting tension threatens the global harmonization of human rights protections in AI.
Jackowski et al. (Thu,) studied this question.