Ex Machina Lex: The Limits of Legal Computability

Abstract

The use of machine learning (ML) to replicate aspects of legal decision making is already well advanced, with various ‘Legal Tech’ applications being used to model litigation risk, and data analytics informing decisions on issues with relevance to law which include probation, predictive policing and credit evaluation. The next step, already being trialled in a number of jurisdictions, will be the use of ML to replicate core functions of legal systems, including adjudication.

We consider the likely consequences of this step using a systemic-evolutionary model of law. From this point of view, many aspects of legal reasoning have algorithmic features which could lend themselves to automation. However, an evolutionary perspective also points to features of legal reasoning which are not consistent with ML: these include the reflexivity of legal knowledge and the incompleteness of legal rules at the point where they encounter the ‘chaotic’ and unstructured data generated by other social sub-systems. We illustrate this point with an example taken from labour law concerning the classification of work relationships.

We argue that the goal of a ‘legal singularity’--a concept advanced by proponents of the use of ML in law--is based on a conception of a functionally complete legal system which, while a mirage, has the potential to divert resources to ultimately fruitless uses, while compromising the autonomy of the legal system and undermining its core modes of operation. Finding the institutional means to maintain law’s system-boundary with technology is an urgent task but one whose success cannot be guaranteed, as there is no principle of societal organisation which guarantees the perpetuation of the rule of law, and the democratic-liberal order it maintains, in the face of current technological changes.

Keywords: Artificial Intelligence, Legal Evolution, Systems Theory, Labour Law, Natural Language Processing, Machine Learning, Computational Linguistics, Fractal Organisation Social Ontology, Legal Singularity

JEL Classification: J01, K10, K12, K31, K34, K40, K41, L50, L51, O30, O31, O32, O33