About | HeinOnline Law Journal Library | HeinOnline Law Journal Library | HeinOnline

LOG IN
LOG IN


26 Auckland U. L. Rev. 92 (2020)
Reason-Giving in the Age of Algorithms

handle is hein.journals/auck26 and id is 92 raw text is: Auckland University Law Review Reason-Giving in the Age of Algorithms JESSICA PALAIRET' Public sector agencies are increasingly using sophisticated machine learning algorithms to assist and make decisions previously made solely by humans. This use of advanced algorithms presents opportunities, but also great risks to administrative law. Many complex algorithms are black boxes , meaning no person can explain how they work. Further, dirty data and machine code can produce discriminatory or biased decisions that are difficult to identify and regulate. The idea that keeping a human in the loop will address these problems is unrealistic and short- sighted. Therefore, this article argues that the use of AI in administrative decision-making necessitates the development of a general duty on decision-makers to provide reasons for their decisions. Administrative law cannot stand still amidst the rise of artificial intelligence. A duty to give reasons is not a silver bullet solution, but it is an essential response to decision-making in the age of algorithms. I INTRODUCTION Over the past 10 years, the use and sophistication of machine learning algorithms have increased at a remarkable rate. We are in the midst of a technological revolution that could either be society's greatest opportunity or its most pressing threat.1 New Zealand's public sector is not immune to this revolution. Public sector agencies are increasingly using artificial intelligence (Al) to make decisions previously made by humans. These decisions range from the automatic approval of Accident Compensation Corporation (ACC) claims to decisions on welfare grants, school funding and immigration.2 These applications of Al are just the beginning. This article focuses on machine learning algorithms that derive rules and predictions from large swathes of data. These algorithms can be used to support human decision- * BA/LLB(Hons), University of Auckland. The author would like to thank Professor Janet McLean QC for her supervision and support, and Matt Bartlett for his help and inspiration. This article was awarded the MinterEllisonRuddWatts Writing Prize for 2020. 1 Yuval Noah Harari Why Technology Favors Tyranny The Atlantic (online ed, New York, October 2018). 2 Stats NZ Algorithm Assessment Report (October 2018) at 36-38. 92 Vol 26 (2020)