There are many ways to use self-learning algorithms in the domain of security and justice, from detecting fraud to solving cold cases. At the same time, we understand all too well how important it is to supervise the algorithms themselves. See, for example, the recent policy briefing Big Data and Security Policies: Serving Security, Protecting Freedom by the Netherlands Scientific Council for Government Policy (WRR).
Should we use algorithms to forecast crime?
An example: the Dutch Police recently featured in the news because of its Crime Forecasting System. The system uses historical data to detect patterns of crime, for example burglaries, muggings and robberies. The police use these patterns to identify where best to deploy policing capacity. This then generates new data and follow-up questions about the efficiency, effectiveness and legitimacy of policing.
What does this mean for the role of the police officer?
The Crime Forecasting System does not side-line professional judgement. Local, contextual knowledge remains important. Detecting crime patterns is not the same as identifying the causes of crime. The Crime Forecasting System does not, for example, make it clear whether action on the part of a housing corporation would be more effective in certain cases than extra policing.
Will we be able to predict every crime in the future?
People adjust their behaviour in the light of new information. Targeted policing can, for example, cause criminals to move on to other victims, times of day, locations or methods. And it is not only real or potential criminals who are impacted by policing, but also numerous other parties, including ordinary people. Behaviour can shift, and it is precisely crime forecasting systems that can help us to remain agile as circumstances change.
Keep doing what we’re meant to do
Having the police use self-learning algorithms does not mean we are moving towards a totalitarian state with omnipresent surveillance. In fact, almost the opposite is true. In an environment that is changing dramatically, it is important for the police to use technology to adapt quickly, so that we can continue to do what we’re meant to do: be there for the people.
These two contrasting images are closely related to the sort of police that we – the people of the Netherlands – want in our society. On the one hand, we want the police to ensure the smooth and orderly operation of the state. In that respect, the police are primarily an instrument of the impersonal state, with the emphasis being on repression. On the other hand, the police are there to support and assist the people by protecting freedom and equality and by precluding the ‘law of the jungle’.