Legal Personality for Artificial Intellects: Pragmatic Solution or Science Fiction?

Abstract

Legal personality defines who 'counts' for the purposes of the law. The impact of artificial intelligence technology provides a unique opportunity to look afresh at this oft-misunderstood legal concept. Artificial intellects, or 'bots', will become an increasingly ubiquitous part of society. These bots will exhibit 'weak AI': seeming to possess intelligence, even though they are not actually conscious. The autonomy and capabilities of these bots will make it increasingly difficult to attribute logically their actions to recognised legal persons. Consequently, legal issues will arise - eg is there an agreement or certainty in a bot-conducted transaction? Applying conventional legal doctrine and/or utilising current legislative initiatives designed to facilitate e-commerce to address these issues proves unsatisfactory. Does legal personality assist? This paper considers the appropriate methodology for assessing legal personality, which turns out to be conditions-based: 'under what conditions does the law treat an entity as a separate nexus of legal rights and obligations?' It proffers and explains a unitary theory of legal personality: legal persons are 'the subjects of legal rights and obligations' or, simply, 'units of legal logic'. Legal personality itself is the relevant legal similarity between the disparate entities recognised as legal persons. An entity's metaphysical nature is irrelevant to legal personality determinations. Legal status is distinct from legal personality, defining the specific legal rights and obligations enjoyed by particular subsets of legal persons. The paper uses corporate personality as a detailed example. It concludes that (a) there is no theoretical barrier to according legal personality to bots; (b) whether or not one should do so is determined by extra-legal (moral, philosophical, historical, economic, political, social, etc) considerations prior to the attribution of legal personality; and (c) that it is too early to conclude that legal personality should be accorded to artificial intellects, although there are already significant pragmatic arguments in favour of doing so.