myAIPMImproving public services using artificial intelligence: possibilities, pitfalls, governance
Paul Henman | Asia Pacific Journal of Public Administration
Abstract
Artificial intelligence arising from the use of machine learning is rapidly being developed and deployed by governments to enhance operations, public services, and compliance and security activities. This article reviews how artificial intelligence is being used in public sector for automated decision making, for chatbots to provide information and advice, and for public safety and security. It then outlines four public administration challenges to deploying artificial intelligence in public administration: accuracy, bias and discrimination; legality, due process and administrative justice; responsibility, accountability, transparency and explainability; and power, compliance and control. The article outlines technological and governance innovations that are being developed to address these challenges.
Introduction
New digital technologies are rapidly changing the landscape for the delivery of public services. Mobile devices teamed with apps bring online public services to wherever the citizen is. Networked and wi-fi technologies enable the provision of information and collection of geo-coded data to be integrated with traditional administrative data, creating “big data” sets for building knowledge about populations and individuals. Automated administrative decision-making processes are being expanded, with artificial intelligence (via machine learning) providing more nuanced ways to make decisions within complex circumstances.
This article focuses on the potential of artificial intelligence (AI) for improving public services. It focuses on three areas: automated administrative decision-making; chatbots; and public governance. The article first provides examples of emerging uses of digital tools and processes and the potential benefits they provide. It then considers some of the technological, ethical, legal and political challenges and pitfalls to such technologies. Finally, it highlights ways in which regulation and governance of AI-based public administration can be enhanced in order to advance its potential, while mitigating its negative consequences to individuals and society.
Before proceeding it is important to clarify terminology. What is meant by “artificial intelligence”? The term artificial intelligence (AI) has been used and enlivened the imaginations of people for decades. Each new wondrous development of computer algorithms and systems – from Weizenbaum’s (1984) ELIZA who engaged people in conversation in the 1960s, and IBM’s Deep Blue program that defeated the world chess master in 1996, to driverless cars of the present – have repeatedly displayed intelligent-like behaviour and challenged our collective thinking about what counts as intelligence. Today, the AI moniker is used as much as hype and marketing as to signal a real shift in how algorithms are being developed. AI is typically used to refer to (systems developed with) machine learning algorithms that self-organise their internal variables and values to achieve desired outcomes. For example, instead of human programmers trying to identify the facial structures that can be used to distinguish between and identify individuals, AI facial mapping systems learn by being trained on large sets of faces and matched faces with names. Current use of AI terminology is often about a specific mode of machine learning, namely artificial neural networks, which are inspired by brain structure and operation. Hardware (and software) developments have led to the recent explosion of AI as they have enabled what previously was not technologically feasible. Hereafter, this article uses the acronym AI to reflect computer systems built (at least partially) using machine learning.
The recent dramatic enthusiasm for AI (as machine learning) and associated plethora of policy discussion papers (Fjeld et al., 2020) signals a fundamental change in digital technologies. In terms of usage in government, it is important to keep in mind that AI is part of a trajectory in digital technologies, systems and algorithms, rather than a fundamental rupture. For example, many of the legal, ethical and social concerns about the use of AI in government are identical to those of non-machine learnt algorithms. The emergence of AI discourse has been beneficial in generating public discussions of how digital technologies are governed to achieve collective principles and aims.
Read more
Improving public services using artificial intelligence: possibilities, pitfalls, governance, Paul Henman, Asia Pacific Journal of Public Administration, 2020