Abstract
Recent advances in automation have slowly begun to infiltrate the global judiciary, sparking controversy regarding the feasibility and desirability of outsourcing judicial decision-making to artificially intelligent systems. Here we surveyed legal academics (n=466) at leading universities across the English-speaking world regarding to what extent they believed artificial intelligence (AI) can, will and should be adopted into the judiciary. Our main findings are threefold. First, participants by-and-large predicted that the majority of judicial decision-making can and will be carried out by AI instead of humans within the next century, and that over one-third could be carried out by AI at the level of humans within 25 years. Second, participants were by-and-large not in favor of outsourcing judicial decision-making to AI, despite believing that doing so would improve the judiciary along dimensions such as fairness, efficiency, predictability, and judicial transparency. Finally, participant responses were significantly associated not only with demographic factors such as politics and gender, but also with jurisprudential factors such as the endorsement of natural law over positivism, suggesting deep connections between legal academics’ views of automating the judiciary and their views of law, morality and politics more generally.
