I gave a presentation on "Predictive analytics and the future of law and regulation" at the Curtin Early to Mid-Career Research Symposium held in Perth on 24 October 2024. The following is the abstract:
With the increasing use of data mining, machine learning and AI by governments around the world, the problems and risks of applying predictive analytics in regulation have gone beyond the realm of science fiction. Predictive policing is reminiscent of Precrime in The Minority Report where people are arrested before they commit crimes. Further, the ignominious Robodebt scheme illustrates the dangers of uncritical deployment of automated decision-making systems in governmental functions. The main aim of this presentation is to explore the conceptual and practical issues with prediction-based regulation by examining the interplay between and among descriptive, prescriptive and predictive rules in law. Understanding how these ‘is’, ‘ought’ and ‘will be’ rules interact with and impact each other are critical for the responsible adoption of both algorithmic technology and predictive regulation by governments. This presentation explains the long-running debate about Hume’s law and whether there can be ‘no ought from is’. It then discusses how the introduction of ‘will be’ rules via predictive analytics further exacerbates the theoretical and regulatory complexity. The presentation sketches out the dynamic qualities and relations of descriptive, prescriptive and predictive rules and what they mean for the future of predictive regulation and law in general.
For more presentations, visit the Presentations page.
Commentaires