Persona ex machina: Artificial intelligence on a spectrum of legal personhood

Abstract

Science fiction has long been fascinated with the potential for sentient, human-like artificial persons. Technological reality has quite some way to develop before an artificial intelligence (AI) approximates these fantasies, and yet individuals, governments and businesses already entrust AI to make automated decisions with potential for real-life impact. As AI capacities increase, we can expect that this treatment will continue to proliferate, increasing also in decision-making complexity and gravity. This thesis aims to gain greater insight into the impact that this shift is already having on legal frameworks, approached from the perspective of legal personhood as a core concept that impacts every branch of the law.

This thesis conceptualises legal personhood on a spectrum, positioning legal persons relative to quasi legal persons and distinct from things. It aims to capture the nuances of legal personhood beyond the implications of a traditional binary understanding of the concept. It will show examples of how legal personhood is subject to change over time from economic, social and pragmatic pressures.

This thesis also attempts to identify the place that AI may already claim on that spectrum, both as its own entity and as a result of increasing utilisation of AI in automated decision-making. Suggestions are also made with respect to an intentional quasi-personhood model for AI, drawing on contemporary and historical examples.