Man and machine: artificial intelligence and judicial decision making
Arthur Dyevre, Ahmad Shahvaroughi
This paper examines how artificial intelligence tools are being used in courts to help judges make decisions about pretrial release, sentencing, and parole. The authors review research across multiple fields showing that while AI risk assessment tools have raised concerns about fairness and transparency, evidence suggests they have surprisingly little actual impact on judges' decisions. The paper calls for more research to understand how judges interact with AI recommendations and what this reveals about both algorithmic and human decision-making.
criminal justiceAI fairnesshuman-AI collaborationjudicial decision-making