Foundations for the future: institution building for the purpose of artificial intelligence governance

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.

Research agenda for using artificial intelligence in health governance: interpretive scoping review and framework

“…While recent technologies, particularly AI, constantly modify human interactions, a fundamental question is: ‘how to ensure digital resilience and collective well-being, while safeguarding liberal democracy and individual rights’ [ 40 ]. To ensure the provision of beneficial answers regarding the mentioned concern, new capacity building in institutions are needed to implement such efforts at national and international levels [ 41 ]. Appropriate technologies, facilities, particular expertise both for users and big data experts [ 42 ], shared guidelines for databases [ 43 ], norms, public relations, and education would remain important factors for the successful adoption of modern technologies.…”

Discerning Between the “Easy” and “Hard” Problems of AI Governance

“…Some example problem domains and literature examples are also given. 2022-01-0013-OA10-TTS [3], [47] Large-scale societal challenges [13] Actors Organizational managers, developers, executive sponsors [11], internal review boards [14], [16] suppliers, auditors, insurance companies [16], professional associations (e.g., IEEE, ACM) [16], investors [48] Supranational legislative bodies, nation-states [30], government agencies [15], [16], new institutions (e.g., European Union oversight agency) [49], global and regional initiatives (e.g., World Economic Forum Global Artificial Intelligence Council), industry-led initiatives (e.g., Partnership on AI) [1], [30], research institutions [16], standardization bodies [50], ecosystems [51], [52]…”

“…Organizational strategies and processes [11], organizational ethics guidelines [32], technical tools and methods (e.g., software engineering workflows, audit trails) [16], [47] Binding legislation [30], human rights frameworks [53], establishment of new global or regional institutions [49] Ethical basis Business ethics, corporate social responsibility [36], [37], [51], deontology [54] Ethics of emerging technologies, data justice, human rights, technology for good [55] Example problem domains Organizational governance processes [11], [12], creation of responsible AI-enabled consumer products [56] Deepfake technologies' undermining of trust in media [41], [42], frameworks for responsible innovation [57] Illustrative literature examples [10]- [12], [31]- [33] [1], [13], [40], [53] The first characteristic, time horizon, is implicit rather than explicit in the literature. Arguably, short-and long-term concerns are involved in both problem areas.…”