Legal authorities are slowly using ML models to predict the outcomes of pending cases, using the corpus of relevant precedent and a case’s fact pattern as inputs.