This dissertation examines how the legal system perceives the introduction of algorithms in courts as a means of handling the demands of time, efficiency, and risk. As technologies like natural language processing and machine learning develop, algorithms can be used to read and process enormous quantities of data, thus opening the possibility of automating tasks once considered solely performed through human thought. Within a social context of hyper-connectivity and accelerating social processes, the judiciary, as part of the legal system, is usually represented as inefficient, too slow and too expensive – and thus a field ripe for the application of technological solutions. In recent years courts have been incorporating algorithms into their functioning to varying degrees, with the Brazilian Judiciary counting among those enthusiastically developing, testing, and implementing in-house solutions to tackle its enormous caseload. Starting from a systems-theoretical standpoint, this monograph observes how the legal system is self-described in the accounts produced by the Brazilian Superior Court of Justice (STJ) and the National Council of Justice (CNJ) as they communicate about their technological initiatives.