AI-related data ethics oversight in UK policing

Abstract

This paper considers the question of how police-related AI projects and data projects in general are normatively assessed in the UK. After locating data ethics in relation to policing ethics, I shall consider the workings of perhaps the leading regional data ethics committee in the UK. I go on to consider the approach of another committee that might in the future provide national data ethics advice for the police. Finally, I summarize the normative ethics frameworks in use in the two committees and their heavy reliance on the concepts of necessity and proportionality. I suggest that these concepts may have to be supplemented by systematic thinking about varieties of harm and the way in which severe harm may generate obligations to prevent it, where prevention may be assisted by AI models.