The criminal justice system has been something of a test-bed for new technologies―witness, for example, the use of CCTV surveillance and DNA profiling technologies. However, with the development of a new generation of smart (AI) tools that can be used to indicate when and where crime is most likely to take place and who is most likely to commit it, reliance on technological solutions to problems of crime and security is set to move to another level. For criminal justice professionals whose concern is to assess the risk of criminality and to manage it, these tools have obvious attractions; but their adoption also has major implications for the kind of society that we want to be.

If AI tools (unlike most criminal justice strategies) actually do ‘work’, then the big question is whether their use for criminal justice purposes is acceptable. Already, many commentators believe that the direction of travel is towards preventive justice and away from traditionally reactive penal justice. Unless communities push back against this trajectory, the availability of AI tools will only reinforce this focus on ex ante prevention of crime. Accordingly, there need to be community-wide discussions about the acceptability of automated justice, of policing by smart machines with humans no longer in the loop, and of false positive and false negative errors.

With a view to opening these discussions, the contributors to the recently published International Journal of Law in Context Special Issue on “Law, Liberty and Technology―Criminal Justice in the Context of Smart Machines,” explore a range of issues, both general and specific. Following an introductory overview (by Roger Brownsword and Alon Harel), Vincent Chiao assesses the prospects for AI tools relative to the complex expectations that we have for the criminal justice system but also measured by the fairly low bar set by current performance, Ben Bowling and Shruti Iyer discuss the issues raised by the use of body-worn cameras/videos by the police, Elizabeth Joh sketches her vision of policing in the smart cities of the future where the regulating technologies are embedded, integrated, and largely privatised, and Stuart Macdonald, Sara Correia and Amy-Louise Watkin examine the use of AI by social media companies who seek to block and remove terrorist content from their platforms. In Europe, we put much of our faith in the latest data protection regimes but, as Orla Lynskey argues, this already problematic body of law is challenged by evolving criminal justice practices that use AI; and, in the final contribution, Deryck Beyleveld and Roger Brownsword argue for a robust set of moral principles to underpin and guide both traditional ex post punishment and technologically-enabled ex ante prevention.

The late Stephen Hawking remarked that AI might prove to be either the best or the worst thing to happen to humanity. While smart tools deployed in the criminal justice system might not be an existential threat, for better or for worse, they might well be game-changers. Communities need to understand which tools do work; but, they also need to debate which uses are acceptable.

Written by Professor Roger Brownsword, King’s College London and Bournemouth University. Read this special issue of International Journal of Law in Context without charge until the end of July 2019.