The Australian Institute for Judicial Administration (AIJA), UNSW Law & Justice, UNSW Allens Hub for Technology, Law, and Innovation, and the Law Society of NSW’s Future of Law and Innovation in the Profession (FLIP Stream) have conducted a joint research project into the use of artificial intelligence (AI) in the courtrooms around the world.
AI has become increasingly used in the legal system due to its efficiency in administrative matters. For example, AI programs can quickly automate e-filing by using data-driven inferences about a defendant to determine sentencing. However, despite this widespread adoption, the report focuses on critical issues arising from the increased use of AI. This includes concerns around computer-based decision resolution software, software-driven logic derivatives, and “AI judges”. The researchers comment on the compatibility of AI with legal values. In the US, the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) tool uses a questionnaire to calculate the likelihood of an offender contravening the law again. In theory, this seems useful, although, the questionnaire is oftentimes opaque, and the tools software poses issues of confirmation bias by the developers. More importantly, the outcome of COMPAS has an influential effect on a judge’s decision on whether an alleged offender can be granted bail or whether an accused offender should be eligible for parole. This already occurred in 2013, when an individual agreed to a plea deal for one year in a county jail for stealing a lawnmower, but COMPAS noted a high risk of reoffending, and the sentencing judge rejected the deal and issued two years in prison. Saving time and resources becomes contrary if automated decisions directly contravene the purpose of the judicial system and moral authority.
That being said, AI has proven itself valuable both in Australia and worldwide in benefiting access to justice. The use of natural language processing in AI permits the conversion of foreign spoken audio to text for judges, witnesses, and counsel. It also minimises human error, particularly in transcription and physical impairments.
AI is certain to integrate itself further with the legal system. What is important for courts to recognise is that AI should be used as a tool or guide that promotes the legal values of open justice, accountability, impartiality, equality, procedural fairness, access to justice, and efficiency.
For a full reading of the report, see here.