Cutting Crime Impact - the human rights and ethics watchdog

Law enforcement agencies (LEAs) are confronted with difficult decisions regularly. They need to decide on effective policing, recidivism of criminals and find ways to improve overall security. Those are complicated decisions which should be taken based on facts and data. This complex collection and analytics exercise is increasingly supported by algorithms assisting LEAs, because they are capable of including more information in the decision-making. However, when automated decisions come into play, a close look at details is needed since lives of citizens and groups are shaped by technology. Oskar Gstrein, Assistant Professor at the Department of Governance and Innovation at Campus Fryslân, works on a Horizon 2020 project called Cutting Crime Impact (CCI) to solve ethical, legal and social issues in automated law enforcement.

What is CCI?
While law enforcement was solely a people’s business a few decades ago, more and more LEAs turn to technology and AI to improve their operations in recent years. A popular example is the widespread use of predictive policing, which uses AI and analytics that is applied to predict the risk of crime certain people and locations are exposed to. Those techniques are used by the police to coordinate how and where to patrol. However, the use of such predictions results in several ethical and legal challenges. The University of Groningen is involved in CCI, an innovative European project coordinated by the University of Salford, to research those limitations. Overall, CCI has two main objectives. First, to reduce and prevent crime and second, to develop an extended European Security Model. Gstrein, who together with his colleagues focuses on legal and ethical implications alike, says the Data Research Centre of Campus Fryslân is a “human rights and ethics watchdog.”

Collaboration with Law Enforcement Agencies
CCI works together with a number of Law Enforcement Agencies from all over Europe, the Dutch, Estonian, Lisbon, Greater Manchester and Lower Saxony State Police as well as the Departament d’Interior of Catalonia. In a collaborative effort with the LEAs, solutions to issues linked to law enforcement and security are designed. The CCI consortium is approached by law enforcement agencies with different questions and problems. In response, the partners work with the agencies to define the baseline goals and design solutions to the initial problems. Many agencies already have systems running and start thinking about innovative ‘toolkits’ mitigating the challenges. There, the CCI consortium assists in the design. “Eventually,” says Gstrein, “the systems look very different to when law enforcement agencies first approached the research initiative: it is still a police idea, but shaped very differently.”

Understanding the product
It is viable to have the expertise in-house and build the right culture in order to understand the systems that are put in place, according to Gstrein: “for a lot of these tools, it is not only important that you can push a button and then it works, but it is also important to understand what is happening.” This is especially important in law enforcement, because every step produces tangible impactful consequences for citizens. Looking back on the history of predictive policing, Gstrein knows what is important: “Only if you have enough expertise in the mix, between disciplines, to build those bridges and understand what the rationality behind the process is, we can design tools supported of AI in a way that we actually get a better law enforcement process in the end.”