Algorithmic Governance: Evaluating the Role of Artificial Intelligence in Public Policy Decision-Making

Abstract

This article discusses the use of artificial intelligence in the public policy and challenges associated with it such as the velocity of change and the impact on democracy. Artificial intelligence is really changing how we do policy. That is creating a tension between the speed at which we can do things and the democratic way of doing things with AI.” States really like intelligence. But scholars who study this do not all agree. There is not one idea here that links the use of algorithms to how things are already done and the checks we have in place. The paper reviews the situation and develops a framework that draws on governance theory, public administration and digital governance to explain how technology, institutions and values interact to shape policy. The framework consists of three components: strong institutions, transparency of algorithms and democratic accountability. Will this interaction between these things allow artificial intelligence to improve policy without losing legitimacy or make problems like bias, secrecy and lack of accountability worse? Other models separate technology and politics; the article provides a lens, which it calls the Policy-Algorithm Governance Matrix, for seeing how different combinations of institutions and algorithms can lead to very different outcomes. This is very useful in practice. It can help guide policymakers and scholars in countries in the Global South that are rapidly digitizing but have weak institutions, which is common. Ultimately they have to ensure that artificial intelligence is deployed in a transparent, accountable and ethical way so that we don’t trade democracy for speed. It is part of artificial intelligence and we have to think about how this plays into policy and democracy.