YU, Tak Ming (2026) A framework for AI governance and public trust in policing. Policing: A Journal of Policy and Practices. (In Press)

PDF (AM to be made available through RR after publication) YuTM_A Framework for AI Governance and Public Trust in Policing_AM.pdf - Accepted Version Restricted to Repository staff only Download (280kB) | Request a copy

Abstract

The Metropolitan Police Service (MPS) has reported high operational performance from the deployment of Live Facial Recognition (LFR). However, incorrect facial recognition matches continue to raise concerns regarding reliability, transparency, accountability, and legitimate AI use. These concerns are important because the public may have a limited ability to understand how AI-enabled systems operate, how decisions are reviewed, and who remains responsible when errors occur. The increasing use of artificial intelligence (AI) in policing raises a broader governance challenge: how new technologies can be used effectively, safely, and accountably while maintaining public legitimacy and trust. Although the AI governance literature shows convergence around transparency, accountability, fairness, and human oversight, different disciplinary traditions emphasise different concepts, mechanisms, and levels of governance, limiting their integration into a coherent structure. This study addresses this disconnection through an integrative literature review. The analysis identified three analytical dimensions: conceptualisations of governance, normative principles, and levels of governance. Further conceptual synthesis identified four recurring governance functions: purpose and value, operational implementation, legal authority and safeguards, and institutional responsibility and oversight. These functions form the conceptual, operational, legal, and institutional layers of the proposed Four-Layer AI Governance Framework. Its application to MPS live facial recognition illustrates how legitimate purposes and boundaries can be connected with operational controls, legal safeguards, human judgement, institutional responsibility, and continuing review. The framework shows why technical effectiveness alone is insufficient to establish legitimate AI use and provides a common analytical architecture for connecting governance concerns across systems and contexts.

Item Type:	Article
Additional Information:	This is a pre-copyedited, author-produced version of an article accepted for publication in Policing: A Journal of Policy and Practices following peer review. The version of record [insert complete citation information provided when published] is available online at: xxxxxxx [URL and DOI of the article will be provided when published].
Subjects:	Computing > Intelligent systems Law and criminal justice > Criminal justice > Policing and criminal investigations
Date Deposited:	24 Oct 2026
Dates:	Date Publication status 22 October 2026 Accepted
School, department or research centre:	School of Human and Social Sciences
URI:	https://repository.uwl.ac.uk/id/eprint/15569
Sustainable Development Goals:	Goal 9: Industry, Innovation, and Infrastructure	Sustainable Development Goals:	Goal 10: Reduced Inequalities	Sustainable Development Goals:	Goal 17: Partnerships for the Goals

Actions (admin access)