AI as a Legal Person?

Abstract

Cyclical advancements in artificial intelligence (“AI”) are usually accompanied by theories advocating the granting of legal personhood to sophisticated, autonomous computers. The paper criticizes such theories as incorrect — a possible result of legal scholars being seduced by incomprehensible technical terminology, sensationalistic stories in the popular press and “creative” photo filters that transform our faces into animals. Discussions as to when computers should be recognized as persons are, logically, outside of the scope of intellectual property law. The granting of legal personhood is not premised on the existence of consciousness, intelligence or creativity. Recognizing an entity as a legal person is a normative choice dictated by commercial expediency, not the result of fulfilling any technical criteria. While it is necessary to acknowledge the blurring of borders between art and (computer) science as well as the increase in the technological sophistication of the tools used by authors and inventors, it is also necessary to state that even an exponential increase in “computer creativity” will not sever the link between the computer and its user. Before discarding the idea of legal personhood for “creative algorithms” once and for all, the paper explores the relationships between autonomy and creativity. In particular, it places technical terms such as “AI” and “autonomy” in their original context and criticizes uninformed attempts to imbue them with normative connotations.