Responsible AI in criminal justice: LLMs in policing and risks to case progression

TL;DR

This paper examines the use of Large Language Models in policing within England and Wales, identifying potential risks and impacts on case progression, and emphasizes the need for responsible implementation practices.

Contribution

It provides a practical framework for assessing risks of LLMs in policing, including specific tasks, risks, and impact examples, guiding responsible AI deployment.

Findings

Identified 15 policing tasks suitable for LLMs

Outlined 17 risks associated with LLM use in policing

Provided 40+ examples of impact on case progression

Abstract

There is growing interest in the use of Large Language Models (LLMs) in policing, but there are potential risks. We have developed a practical approach to identifying risks, grounded in the policing and legal system of England and Wales. We identify 15 policing tasks that could be implemented using LLMs and 17 risks from their use, then illustrate with over 40 examples of impact on case progression. As good practice is agreed, many risks could be reduced. But this requires effort: we need to address these risks in a timely manner and define system wide impacts and benefits.