The use of Agentic AI in judicial decision-making

AI has the potential to significantly accelerate processes traditionally reliant on expert human input, which are often time-consuming and labour-intensive. A notable example is the judicial system, where cases can take years to resolve, and the high costs of court proceedings frequently deter individuals from pursuing justice. However, integrating AI into judicial processes presents substantial ethical challenges for obvious reasons. Indeed, given the potentially life-altering consequences of a judgment, ethically testing AI-assisted judgments in real-world setting is extremely challenging. This project aims to address the ethical challenges around testing AI-assisted judgments through the use of Early Neutral Evaluations (ENE). ENE is an alternative dispute resolution (ADR) process most commonly used in commercial law that helps parties assess the merits of a case at an early stage, to understand the likelihood of success, avoid unnecessary litigation and preserve relationships. ENE lends itself to this type of research as it is non-binding on the parties and provides a testbed with less risk. There is growing recognition of the benefits of ENE across other court business areas and jurisdictions.