Rubius taught artificial intelligence to look for patterns in the work of judges, to evaluate the effectiveness of lawyers and even to predict the outcome of cases.
Our system analyzed 90 399 court decisions of Tomsk judges for 9 years. As the result, we received the rating of the most effective lawyers and several interesting patterns. For example, female judges tend to impose penalties and more often find defendants guilty of traffic violations, deforestation and economic crimes. Male judges are more likely to convict of beatings and home invasions. But in cases of a power abuse or an unauthorized absence the sentence does not depend on judge's gender.
Such an judicial analysis also may help to search for precedents, identify judicial errors and predict prejudiced attitude to defendants.