Legal problems are prediction problems. Lawyers need to predict how the law applies to new situations to better advise clients; judges need to predict the consequences of their decisions. Better predictions make for a better legal system. Dr. Niblett researches how artificial intelligence can make better legal predictions with machine learning, showing how computers can "think" like a lawyer. He investigates the opportunities and challenges of using artificial intelligence to improve legal decision-making. His research explores how algorithms will change the structure of law, forcing us to address the (ultimately human) question: what exactly do we want the law to do?