PhD defense Joshua Brand: Explaining our way back to humanity: Reclaiming the foundations of the moral and positive law in the age of AI decision-making through explainability

Jury

• Filippo Santoni de Sio, Professor, Eindhoven University of Technology (Reviewer)
• Alexei Grinbaum, Directeur de recherche, CEA Saclay (Reviewer)
• Emily Sullivan, Senior Lecturer, University of Edinburgh (Examiner)
• Carina Prunkl, Researcher Scientist, Inria, University of Oxford (Examiner)
• Winston Maxwell, Professor Emeritus, i3, CNRS, Télécom Paris (Supervisor)
• John Zerilli, Senior Lecturer, King’s College London (Co-supervisor)

Abstract

Artificial intelligence (AI) has increasingly become an omnipresent and opaque force in the value judgements and decisions that directly bear on our lives, doing so in ways that we do not always understand. Current literature responds with the argument that AI ought to be explainable, at least in contexts with moral and/or legal weight, to ensure transparency, trustworthiness, accountability, and empowerment of decision subject agency. Making use of contemporary Kantian and Aristotelian theories, this thesis does not disregard this approach outright. Rather, it goes further by arguing that explainability has a deeper value: sustaining the very structure of the moral and positive law through the principle of reciprocity—the duty- right relationships upon which all other considerations and outcomes, including efficiency and accuracy, depend. While this results in a far broader and stricter implementation of explainability, it makes for clearer moral and policy intervention.