Artificial intelligence and the model of rules: better than us?

Artificial intelligence (AI) decision-making systems are already being extensively used to make decisions in situations where legal rules are applied to establish rights and obligations. In the United States, algorithmic systems are employed to determine the rights of individuals to disability benefits, to evaluate the performance of employees, selecting who will be fired, and to assist judges in granting or denying bail and probation. In this paper I explore some possible implications of H. L. A. Hart’s theory of law as a system of primary and secondary rules to the ongoing debate on the viability and the limits of an adjudicating artificial intelligence. Although much has been recently discussed about the potential practical roles of artificial intelligence in legal practice and assisted decision making, the implication to general jurisprudence still requires further development. I try to map some issues of general jurisprudence that may be consequential to the question of whether a non-human entity (an artificial intelligence) would be theoretically able to perform the kind of legal reasoning made by human judges.