Machine Learning Algorithms in Administrative Decision-Making

Paola Savona 
(Department of Law, Lumsa University, Italy)

Abstract: Machine learning algorithms play a significant role in the digital economy. They suggest products and services to clients, select friends and news, give navigation advice to drivers, make translations. Moreover, learning algorithms are increasingly used to make important decisions about individuals. Companies, for example, rely on machine learning to approve loan, evaluate investments, calculate insurance risks, evaluate workers’ performance or select people to hire. Governments use it to detect terrorists and prevent future attacks, target citizens or places for police scrutiny, select tax payers for audit, detect frauds, grant or deny visas, and more. The influence of machine learning in administrative decision-making might rapidly grow in the near future. The paper analyses opportunities and risks involved in relying on learning algorithms to support or to make administrative decisions with the aim of understanding the challenges that the use of those tools poses to the core principles of the rule of law.

Key words: algorithms; machine learning; data mining; big data; administrative decision-making; black box

JEL codes: K230