Artificial Intelligence (AI)

Five AI management strategies—and how they could shape the future

Atlantic Council GeoTech Center AI Connect II

2025

Co-authors: Wendy Hall

Themes: Artificial Intelligence (AI)

Category: Blog, Open access

The AI governance regime is evolving, and fortunately it is focused on predictable or evident risks, not speculative existential threats. Governments can legislate, and some have. New institutions, such as the EU’s AI Office and Britain’s AI Safety Institute, have emerged. Supranational groupings foster cooperation and standards, such as the United Nations AI Advisory Body, or the Group of Seven’s Hiroshima Process, and alongside these has been a tsunami of summitry and experience sharing. The combination of government regulation, global policy frameworks, research and testing infrastructure, and best practices will gradually coalesce into a recognizable AI governance regime with established norms and shared principles. In this shuffle, we see the repurposing of the ideal types of governance of the O’Hara/Hall Four Internets framework as governance strategies in the AI context, which we term Artificial Intelligence Management Strategies, or AIMS.

Five AIMS: Lessons from Internet Governance for Artificial Intelligence Management Strategies

Sciences Po Tech & Global Affairs Innovation Hub

2025

Co-authors: Wendy Hall, Pierre Noro

Themes: Artificial Intelligence (AI)

Category: Open access, White paper/report

To help scaffold coherent, coordinated, and enforceable rules and institutions, Dame Wendy Hall, Kieron O’Hara, and Pierre Noro reinterpret the Four Internet models elaborated by Hall and O’Hara in their influential book Four Internets: Data, Geopolitics, and the Governance of Cyberspace in regard to AI technologies. This translation, grounded in an analysis of the historical, socio-economic, and ideological differences distinguishing the context that shaped Internet governance and the current one, yields many enlightening insights and is the foundation of five Artificial Intelligence Management Strategies (AIMS). With many illustrations to exemplify their core tenets, their limits and their intersections, this paper offers the Five AIMS as cardinal concepts to help AI governance stakeholders, especially public and private decisionmakers, navigate the upcoming AI Action Summit and future governance conversation. Concluding on a set of ongoing research questions reflecting open policy challenges, it is a foundational step towards cementing the Five AIMS as a suitable framework for understanding the governance of AI.

Zuckerberg’s cave: smartness and discipline in digital modernity

Dariusz Brzezinsky, Kamil Filipek, Kuba Piwowar & Malgorzata Winiarska-Brodowska (eds.), Algorithms, Artificial Intelligence and Beyond: Theorizing Society and Culture of the 21st Century, 72-88

2024

Themes: Artificial Intelligence (AI), Digital modernity

Category: Book chapter, Peer reviewed

This chapter discusses some of the narratives or imaginaries that help structure our experience of, and the development of, the technological suite that mediates the affordances of social life in wealthy Western societies (and increasingly in less wealthy ones, and elsewhere than the West) in the 2020s. It argues that, far from society having transcended modernity to become postmodern, modernity has evolved, from the ‘high’ or ‘reflexive’ modernity of the late 20th century to a characteristic digital modernity. The chapter draws out some of its consequences, sketching its four institutional dimensions. Its disciplinary stance towards individuals gives rise to a myth which is termed Zuckerberg’s Cave, which is compared to the pre-modern cave of Plato, as well as analogous myths of modernity. The realistic prospects of digital modernity are assessed, setting out both its practical risks and its attractions to citizens and policymakers alike.

Explainable AI and the philosophy and practice of explanation

Computer Law and Security Review, 39, 105474

2020

Themes: Artificial Intelligence (AI)

Category: Journal article, Peer reviewed

Considerations of the nature of explanation and the law are brought together to argue that computed accounts of AI systems’ outputs cannot function on their own as explanations of decisions informed by AI. The important context for this inquiry is set by Article 22(3) of GDPR. The paper looks at the question of what an explanation is from the point of view of the philosophy of science – i.e. it asks not what counts as explanatory in legal terms, or what an AI system might compute using provenance metadata, but rather what explanation as a social practice consists in, arguing that explanation is an illocutionary act, and that it should be considered as a process, not a text. It cannot therefore be computed, although computed accounts of AI systems are likely to be important inputs to the explanatory process.

AI in the UK: A Short History

Wendy Hall & Jérôme Pesenti, Growing the Artificial Intelligence Industry in the UK, 18-20

2017

Themes: Artificial Intelligence (AI)

Category: Book chapter, Open access

A brief narrative of the major events in British AI research from Alan Turing to DeepMind.

Avoiding omnidoxasticity in logics of belief: a reply to MacPherson

Notre Dame Journal of Formal Logic, 36(3), 475-495

1995

Co-authors: Han Reichgelt, Nigel Shadbolt

Themes: Artificial Intelligence (AI)

Category: Journal article, Open access, Peer reviewed

In recent work MacPherson argues that the standard method of modeling belief logically, as a necessity operator in a modal logic, is doomed to fail. The problem with normal modal logics as logics of belief is that they treat believers as “ideal” in unrealistic ways (i.e., as omnidoxastic); however, similar problems re-emerge for candidate non-normal logics. The authors argue that logics used to model belief in artificial intelligence (AI) are also flawed in this way. But for AI systems, omnidoxasticity is impossible because of their finite nature, and this fact can be exploited to produce operational models of fallible belief. The relevance of this point to various philosophical views about belief is discussed.