A Research Network to Explore the Use and Regulation of Predictive Analytics in Policing

Advances in big data, information processing technology, and artificial intelligence are purported to confer immense benefits to law enforcement agencies when investigating crimes and administering criminal justice, including by enabling more efficient allocation of resources; speedier apprehension of criminal suspects; and, providing new methods of assessing reoffending risk. However, the adoption of such technologies poses new problems for those concerned in ensuring that the criminal justice system is fair and effective. Predictive analytics have the potential to enable the state to engage in population management and control on an unprecedented scale. For example, algorithmic tools are increasingly being used to take data points about an individual that are relatively benign in isolation (such as age, address, sex, education) and use this information to develop individual risk scores, which lead to further social sorting or intervention. The dangers of such a shift in the balance of power between the state and the citizen are nowhere more pronounced than in law enforcement, where the coercive arm of governmental power is brought to bear on the citizen. In the absence of robust and clear laws to moderate the exercise of this power, and ensure accountability and transparency, the use of predictive policing tools could have significant adverse consequences to fundamental human rights and democratic norms such as the rule of law.

The increasing use of advanced algorithmic technologies for predictive policing raises pressing questions, such as what are the potential benefits and pitfalls of these practices? What unintended impacts and consequences arise from their use? How can, and should, the law respond - are existing legal mechanisms adequate, and if not what new forms of regulation are required? What will be the impact of the UK's departure from the frameworks developed, and what position should the UK adopt in respect of emerging legislative proposals at the EU level? We contend that expertise crossing legal disciplinary boundaries is necessary to answer these questions and effectively understand, regulate and safeguard against the myraid risks of predictive policing. In addition, there is an urgent need to bridge the gap that currently exists between academics and policymakers working in this area.

The proposed Predictive Policing Network (PPN) will be the first dedicated network concerned with developing collaborative opportunities for legal researchers, regulators and policymakers to explore these questions. It aims to foster dialogues between those who share a concern in ensuring that any use of predictive technologies and other artificial intelligence in policing is guided by democratic principles and subject to effective regulation. The Network will draw on the varied expertise and research of Network members to establish an academic community and hold a series of events that break down the existing barriers between different legal disciplines and between academic researchers, regulators and policymakers on the European continent. In doing so it will enable the opportunities and challenges presented by predictive policing to be better understood and appraised, aiding the development of a responsible AI and advanced algorithmic ecosystem in UK criminal justice policy and practice.