Ethox Seminar: Automated prejudice: Should algorithms assess risk in the criminal justice system?
In the criminal justice system, risk assessments of an individual’s propensity to offend or reoffend are made to determine what is ‘acceptable risk’; where individual deprivation of liberty prevents future societal harm. Traditionally, these decisions have been made by humans, assisted by an in-depth knowledge and clinical assessment of the individual, true to the principle of individualised justice. However, in recognition of prejudice in human decision-making and a desire for accuracy and expediency in risk assessments, algorithms now appear at every stage of the criminal justice system. But can these tools ever offer us a judgement that is transparent and free from prejudice, or are we simply automating bias against marginalised groups in our society?
In this talk, I’ll discuss the ethical implications of risk assessment tools currently being used to make decisions in the criminal justice system and consider the ‘ethical debt’ to society created by the rapid development of tools which are disadvantaging minoritised groups.

Joining link: medsci.zoom.us/j/91251667656
Date: 3 October 2024, 14:30 (Thursday, -1st week, Michaelmas 2024)
Venue: Richard Doll Building, Old Road Campus OX3 7LF
Venue Details: Lecture Theatre and on Zoom
Speaker: Hope Kent (University of Exeter)
Organising department: Ethox Centre
Organiser contact email address: admin@ethox.ox.ac.uk
Part of: Ethox Centre Seminars
Booking required?: Not required
Audience: Public
Editors: Aileen Mooney, Isaac Cheng