No abstract

In the development of governmental policy for artificial intelligence (AI) that is informed by ethics, one avenue currently pursued is that of drawing on “AI Ethics Principles”. However, these AI Ethics Principles often fail to be actioned in governmental policy. This paper proposes a novel framework for the development of ‘Actionable Principles for AI’. The approach acknowledges the relevance of AI Ethics Principles and homes in on methodological elements to increase their practical implementability in policy processes. As a case study, elements are extracted from the development process of the Ethics Guidelines for Trustworthy AI of the European Commission’s “High Level Expert Group on AI”. Subsequently, these elements are expanded on and evaluated in light of their ability to contribute to a prototype framework for the development of 'Actionable Principles for AI'. The paper proposes the following three propositions for the formation of such a prototype framework: (1) preliminary landscape assessments; (2) multi-stakeholder participation and cross-sectoral feedback; and, (3) mechanisms to support implementation and operationalizability.

This white paper was originally published by the World Economic Forumand is re-published with permission. Council on Neurotechnologies, which is exploring the ethical principles of using data and technology to improve global mental health and well-being, is made up of the following individuals:

Recent progress in artificial intelligence (AI) raises a wide array of ethical and societal concerns. Accordingly, an appropriate policy approach is urgently needed. While there has been a wave of scholarship in this field, the research community at times appears divided amongst those who emphasize ‘near-term’ concerns and those focusing on ‘long-term’ concerns and corresponding policy measures. In this paper, we seek to examine this alleged ‘gap’, with a view to understanding the practical space for inter-community collaboration on AI policy. We propose to make use of the principle of an ‘incompletely theorized agreement’ to bridge some underlying disagreements, in the name of important cooperation on addressing AI’s urgent challenges. We propose that on certain issue areas, scholars working with near-term and long-term perspectives can converge and cooperate on selected mutually beneficial AI policy projects, while maintaining their distinct perspectives.

No abstract

Governance efforts for artificial intelligence (AI) are taking on increasingly more concrete forms, drawing on a variety of approaches and instruments from hard regulation to standardisation efforts, aimed at mitigating challenges from high-risk AI systems. To implement these and other efforts, new institutions will need to be established on a national and international level. This paper sketches a blueprint of such institutions, and conducts in-depth investigations of three key components of any future AI governance institutions, exploring benefits and associated drawbacks: (1) “purpose”, relating to the institution’s overall goals and scope of work or mandate; (2) “geography”, relating to questions of participation and the reach of jurisdiction; and (3) “capacity”, the infrastructural and human make-up of the institution. Subsequently, the paper highlights noteworthy aspects of various institutional roles specifically around questions of institutional purpose, and frames what these could look like in practice, by placing these debates in a European context and proposing different iterations of a European AI Agency. Finally, conclusions and future research directions are proposed.

No abstract

Recent years have seen an increase in artificial intelligence (AI) capabilities and incidents. Correspondingly, there has been an influx of government strategies, panels, dialogues and policy papers, including efforts to regulate and standardize AI systems [12,20,37,52]. A first step in most of these efforts is to delineate the scope of the resulting document, typically by either outlining a range of standard technical definitions of AI [76,85] or referencing existing scholarly work [73]. After defining their scope, many policy documents published by governments delve deeper into the 'type' of AI they wish to solicit from industry players and deploy nationally or globally. This largely serves to ensure that the strategies, policy discussions and AI-related milestones sketched within these documents are guided by a 'north star' , or overarching goal. The north star should be comprehensible to all who read and implement the document. Describing the north star allows a non-technical audience to follow and partake in the relevant policy discussions, though it does not replace technical definitions. Although more could be said as to why this is being done and whether it is sensible, such discussion is outside the scope of this paper. Instead, I focus on and contextualize some of these 'north star' definitions themselves. In particular, I explore one of the most prominent recent descriptions: the EU's concept of "trustworthy AI. " I explain its background, its international effects and its drawbacks in more depth. What is in a name? What is in "trustworthy AI?".