ABSTRACT
Over the past couple of years, analysis of evidentiary reasoning has taken centre stage in the artificial intelligence (AI) and law community, after decades of almost exclusive preoccupation with formalizing and modelling legal interpretation and reasoning with rules.2 For many reasons, this shift in emphasis is highly welcome. In legal practice, the overwhelming majority of cases that reach the courts are decided on questions of fact. Academic ‘highcourtitis’, the obsession with a statistically small number of appeal court cases as basis for theories of legal reasoning, distorted the design philosophy of legal expert systems in similar ways as the sometimes distorted picture such renowned sociologists as Max Weber drew of the law. From this perspective, the past preoccupation of legal AI to develop realistic models of legal interpretation was trying to solve an overly complex problem with little practical utility.
