Identification, Categorisation and Forecasting of Court Decisions

Masha Medvedeva’s PhD dissertation ‘Identification, Categorisation and Forecasting of Court Decisions’ focuses on automatic prediction and analysis of judicial decisions. In her thesis she discusses her work on forecasting, categorising and analyzing outcomes of the European Court of Human Rights (ECtHR) and case law across Dutch national courts. Her dissertation demonstrates the potential of such research, but also to highlight its limitations and identify challenges of working with legal data, and attempts to establish a more standard way of conducting research in automatic prediction of judicial decisions. Medvedeva provides an analysis of the systems for predicting court decisions available today, and finds that the majority of them are unable to forecasts future decisions of the court while claiming to be able to do so. In response she provides an online platform JURI Says that has been developed during her PhD, and is available at jurisays.com. The system forecasts decisions of the ECtHR based on information available many years before the verdict is made, thus being able to predict court decisions that have not been made yet, which is a novelty in the field. In her dissertation Medvedeva argues against ‘robo-judges’ and replacing judges with algorithms, and discusses how predicting decisions and making decisions are very different processes, and how automated systems are very vulnerable to abuse.