You are here

PP: AI Beats Out Supreme Court Experts

Submitted by eriklee on Fri, 05/05/2017 - 01:07

Researchers designed a new algorithm that predicts Supreme Court decisions with pinpoint accuracy. The algorithm draws from data from the Supreme Court Database, which contains records dating back to 1791. The algorithm compiles information between 1816 and 2015 and analyzes factors such as judges, votes, term, issues, and court circuits. To test the algorithm, it sequentially goes through the years, starting at 1816, and tries to make a prediction based on the previous year’s data. The algorithm was able to predict the results with 70.2% accuracy. Well qualified experts have 66% accuracy. This can be big as court decision predictions can help the plaintiffs chose which courts to appeal, companies can choose how to handle cases, lawyers can determine how to proceed with the case, and appellants in high profile cases can decide whether to pursue the case. The researchers hope to factor in expert and public statistics and opinions to make the predictions more precise.