Sentencing and Artificial Intelligence

Described as the ‘first collective work devoted to the ethical and penal theoretical considerations of the use of AI at sentencing’, this volume takes us on a deep dive into the world where AI influences (and sometimes determines) sentencing decisions. Where algorithms are used to predict a person’s likelihood of reoffending and where decision trees nudge sentencers to one particular disposal over another.

The law and its practitioners are not, traditionally, the most welcoming when it comes to embracing new technology. With this sentiment in mind, the idea of relinquishing the responsibility of sentencing wholly to AI is one that is, understandably, met with significant scepticism.

In this book, the writers explore commonly raised ethical considerations. Chapters dedicated to the transparency of AI tools discuss just how much of the programming should and needs to be revealed to those on the receiving end of an AI-generated decision. Chapters on fairness grapple with whether AI could serve as an objective tool that reduces instances of unconscious bias, prejudice and discrimination that humans display in their judicial decision-making.

The answer is no. Simply put, datasets used to create AI tools contain decisions that are prejudiced and discriminatory – and the risk remains that AI tools will simply echo these in its output. For this reason, extreme scrutiny and regulation is required to monitor outcomes and facilitate trust in the justice system.

This book includes several interesting studies and cases from the US on the use of AI in sentencing. While we in England and Wales are not utilising AI in sentencing to the same degree, the results from studies discussed raise the question of how much is actually known about the risk assessment tools used in our jurisdiction (the Offender Group Reconviction Scale in pre-sentence reports and the Offender Assessment System when considering sentence plans post-conviction and parole). The NGO Fair Trials is leading on research and campaigns on this topic in Europe and is calling for a ban on risk-assessment systems for the harms they produce.

How far we will go to embrace AI in sentencing remains to be seen. This volume highlights the importance of critical assessment before further adoption.

Kitan Ososami is a pupil barrister at Red Lion Chambers, London