Towards a Robotic Justice: Problems and Perspectives of the Use of Artificial Intelligence in Judicial Systems

In my talk, I will investigate benefits, risks and prospects of the use of AI in courts, in particular in the decision-making phase. After a brief explanation of the difference between “expert systems” and “machine learning”, I will distinguish between two types of AI tools currently used in the field of justice: predictive case law analysis tools, which manage sentences through machine learning techniques, and risk assessment tools, used to predict future criminal behaviors, which are mostly diffused in the North American judicial systems. I will argue that while the first ones are essentially unable to substitute the judge’s decision-making, as this would illogically lead to substitute a “decision” with a “prediction”, the second ones instead are suitable for their intended function, namely, to make statistical predictions about future facts.