For certain court cases, computers can autonomously arrive at (draft) decisions, and for virtually all court cases, they can play an important supporting role for the judge and the trial lawyer. But although artificial intelligence can play an important efficiency- and quality-enhancing role, there remains a clear added value of the individual judge in speaking justice. This is the conclusion of mr. Manuella van der Put in her doctoral research, which she defends at Tilburg University on Wednesday, May 25, 2022.
Judges in the Netherlands are experiencing an increasing workload, which also puts pressure on the judiciary as a whole. To reduce the workload, maintain and improve the quality of the judiciary, artificial intelligence (AI) might offer a solution. For example, KI is able to quickly process large amounts of information, recognize and describe patterns, all parts of the work of judges and trial lawyers. And in part, it already does.
Manuella van der Put, senior judge at the District Court of East Brabant, focused her doctoral research specifically on the process of judicial adjudication to explore the potential of KI for the judiciary. To do so, she used literature review and multidisciplinary focus groups with both legal experts and experts in KI, among others. After careful analysis, she had specific design and quality criteria used to design an artificially intelligent system for judging minor traffic violations, known as Mulder cases. She then tested and evaluated that system. Incidentally, the system has not yet been implemented.
Artificial intelligence can play a role in the process of judicial adjudication in several areas, Van der Put concludes from the study. For certain (still only simple) court cases, computers can arrive at decisions autonomously, and for almost all court cases it can play an important supporting role for even greater efficiency and quality improvement.
For the relatively simple Mulder cases, for example, the system could extract facts from a case, check information such as the appeal deadline and collateral (has a fine been paid or not), analyze a case by linking the information to similar cases in the database, and suggest a decision based on the most common outcome in similar cases. The information this provided saved time and provided a broader understanding of case law. Moreover, after further development, the system should be able to justify that proposed decision based on the motivation from previous cases, and propose a draft judgment using text blocks and formats.
The system was not yet able to independently consider relevant law articles and case law when making the final decision in an individual case, nor to legally qualify the facts. The judge and artificial intelligence cannot do without each other, Van der Put concludes. There remains a clear added value of the individual judge in speaking law, but AI can play an important efficiency- and quality-enhancing role.
The judiciary should certainly embrace artificial intelligence, Van der Put said. It is important to (continue to) see how it can contribute to the challenge of doing justice for and to individual citizens and our society.
