Algorithmic Fairness

Across jurisdictions, criminal justice systems are enamored with the evidence-based practices movement. The idea is to utilize the best scientific data to identify and classify individuals based on their potential...

Prof. Melissa Hamilton presents her research exploring forensic AI’s role and potential flaws in criminal risk assessments within justice systems.

The webinar begins with a speech by Dr Alex Leveringhaus (Surrey Politics) on the role of AI technologies in influencing democratic processes; in particular, elections. This is followed by Prof Amelia Hadfield (Surrey Politics) who talks about how our policy-making processes have become more and more data-driven, and how this raises important questions on where to draw the line on AI affecting political decision-making. Then, Dr Adam Wyner (Swansea Computer Science) underlines a common misconception in picturing AI as solely consisting in Bottom-Up machine-learning mechanisms. Instead, the actual picture is one where Top-Down approaches to knowledge engineering are as relevant as ever for dealing with complex datasets. Finally, Mr Miklos Orban (OPL/Explico) explains how there is very little understanding in society as to the purposes, capabilities and implications of AI. In fact, there seems to be a tendency to polarize the discussion in two extremes, none of which captures correctly the promises as well as the concerns involved with AI.

The ProPublica/Northpointe racial bias debate, and the broader issue of algorithmic fairness, present significant dilemmas for criminal justice officials, legal practitioners, data scientists, and policymakers. This...