The use of Artificial Intelligence in the Dutch courtroom

Abstract

With the rise of data, many sectors have been implementing a form of automatising in recent times. This brings a lot of possibilities. As a matter of fact, an automated system is – unlike the human brain – not dependent on food nor sleep and less prone to errors. The efficiencies that could be gained by ‘working’ 24 hours, 7 days a week without a single moment of pause is enormous. An Artificial Intelligence system (hereinafter: AI system) could provide a relief for many sectors. In the literature, the application for an AI system within the judicial system has been advocated for. Nevertheless, the judicial system, especially regarding the rights of a suspect, needs many more considerations than the workload-relief argument alone. In this thesis, the possible implementation of AI – or a robot judge – within the Dutch judicial system will be explored. In this research, a special focus will be given to the rights of a suspect as set out in article 6 of the European Convention on Human Rights (hereinafter: ECHR). Article 6 ECHR provides the right of a fair trial to a suspect, where relevant, with the following provision: “In the determination of his civil rights and obligations or of any criminal charge against him, everyone is entitled to a fair and public hearing within a reasonable time by an independent and impartial tribunal established by law”. From clause 1 of article 6 ECHR, the following components can be extracted: a fair, independent and impartial tribunal. Implicitly, article 6 ECHR also provides the right to an explanation on how this tribunal reached to its verdict. In practice, this is also referred to as the right to transparent verdicts. With the use of the research question “to what extent can Artificial Intelligence be used in the Dutch courtroom while still maintaining the rights for the parties involved, especially as how has been set out in article 6 ECHR?”, an answer is formulated on the question whether the Dutch judicial system is ready for – and capable of – the next judge: a robot judge. This research uses a systematic literature review as a starting point. With such review, information is gathered from both technical as well as legal point of views to answer each sub-question individually. This research also specifically looks at the way other countries have implemented or created a form of an AI tool in and as court. Next to this, four interviews have been held with potential stakeholders as well as persons of interest. The reason why these interviews are held is to confirm and check what has been found from the literature assessment as well as to gain new (internal) insights on this topic. As an answer to this research question, an AI system can be implemented in the Dutch courtroom if it consists of two (main) components: a linguistic system and a machine learning system. Additional rules to use this robot judge are proposed as well, for example that continuous feedback-loops (of the probabilities for example) should be built in and that the system should be continuously monitored (and adjusted) for potential biases. The robot judge can both function as and in court. Additionally, a robot judge can sit at the table of the Dutch Council for the Judiciary and their ambitions to (further) litigate digitally and publish more, if not all, cases online.