Managing Algorithms for Public Value

Abstract

Public organisations increasingly rely on machine learning algorithms in performing many of their core activities. It is therefore important to consider how algorithms are transforming the public sector. This article aims to clarify this by assessing algorithms from a public value perspective. Based on a discussion of the literature, it is demonstrated that algorithms are generally expected to strengthen organisational performance on a first cluster of values related to the ability to be effective and efficient (sigma values). At the same time, the use of algorithms is linked to negatively affect a second cluster of values that involves fairness and transparency (theta values). In the current academic debate, little attention is given to an important third cluster of values: the ability of organisations to be adaptive and robust (lambda values). This discussion highlights that algorithms invoke public value opportunities, but also public value risks and trade-offs. This article therefore presents five principles for managing algorithms from a public value perspective.