A 'biased' emerging governance regime for artificial intelligence? How AI ethics get skewed moving from principles to practices

Over the past few years, the awareness that the full potential of artificial intelligence (AI) could be attained only through the establishment of a trustworthy and human-centric framework has expanded, thereby prompting demand for regulatory frameworks as well as engendering a flourish of initiatives that set ethical codes and good governance principles for AI development. This study investigates whether the convergence of many of the proposed ethical frameworks around a narrow set of values and principles may be interpreted as a case of transnational norms emergence, a pre-condition for a more structured global regulatory framework or policy regime. Moreover, it explores how this emerging normative framework is reframed in its concrete implementation. Findings suggest that AI governance poses a complex dilemma: while its hybrid governance ecosystem entrusts developers and deployers, mainly from the private sector and technical communities, with the task of translating principles into workable tools, their institu-tional logics substantially narrow the scope and purposes of the ethical approach.

Automated summary

In recent years, the need for a trustworthy and human-centric framework for artificial intelligence (AI) has led to the emergence of regulatory frameworks and ethical codes. This study explores whether these ethical frameworks can be seen as the emergence of transnational norms and if they can contribute to a global regulatory framework for AI. The study also highlights the challenges in implementing these ethical frameworks within the current institutional logics.