Artificial Intelligence, policing and ethics - a best practice model for AI enabled policing in Australia

Abstract

The application of Artificial Intelligence (AI) to policing processes and practices has transformative potential. Police departments around the world have begun deploying AI to assist with various tasks and processes. Examples of AI applications for policing include automated license-plate readers, crime prediction software tools that forecast when, where and who might commit or be the victim of a crime, and tools that simulated environments and scenarios with the objective of training police officers or exploring arguments and outcomes in criminal cases.

Despite its potential, utilising AI for policing also comes with risks and challenges. A significant risk in the application of AI to policing is that of hidden bias and the injustices and inequalities this bias may produce. Another challenge is ensuring that algorithms are well understood and can be correctly interpreted by the police departments using them. Finally, it is important that the use of AI by police has broad community support, as social capital and public trust are well established and essential components of a sound policing framework.

The objective of this article is to provide a starting point for the development of a best practice model for the application of AI to policing in Australia. To achieve this objective, the article begins by introducing the concept of AI. It then maps a range of applications for AI in the context of policing and law enforcement. This mapping process helps to demonstrate the potential of AI for policing, as well as the risks and challenges it presents. The article continues by exploring some of the key risks and challenges in greater depth, drawing on experiences with the use of AI in policing contexts around the world. It then applies an ethical lens to the analysis of these experiences and concludes with a draft framework for a best practice model for the application of AI to policing in Australia. Such a best practice model would be the first of its kind and could put Australian police departments at the forefront of AI application – enabling Australian police to deploy AI in a way that has broad stakeholder support, maximising effectiveness and ensuring ethical concerns are adequately addressed.

Keywords: Artificial Intelligence; Policing; Law Enforcement; Ethics; Governance; Bias; Law; Technology