Could algorithms help magistrates and judges in making sentencing decisions?

Dr Nigel Stobbs

QUT Faculty of Law’s Dr Nigel Stobbs is collaborating with a team of legal experts, judges and data scientists to create a framework for the efficient and ethical use of machine learning algorithms and artificial intelligence in Australia’s criminal justice system.

Together with Professors Dan Hunter and Mirko Bagaric from Swinburne University, Dr Nigel Stobbs recently published details of a proposed trial of sentencing algorithms that will provide magistrates and judges with historical and predictive data to assist in making sentencing decisions.

The current stage of the project involves working with stakeholders to determine what levels of functionality, transparency and control are required over forecasting outcomes, to overcome the well-established risks of ‘Algorithmic Aversion’, which make users less confident in the value of machine learning tools.

The team is seeking funding for the next phase of the project in 2019, which will include development of a beta algorithm for field testing with judges in Queensland and Victoria and evaluation of user preferences for data visualisation interfaces and desktop dashboards.

Nigel will be presenting preliminary results of this project within the proceedings of the International Society for Therapeutic Jurisprudence, held in conjunction with the XXXVIth International Congress on Law and Mental Health in Rome, 21–26 July 2019.

To find out more, read the publication Can sentencing be enhanced by the use of artificial intelligence?