Predictive Policing: Keeping Tabs on Society

November 5, 2015 | 23:19
Predictive Policing: Keeping Tabs on Society
Predictive Policing: Keeping Tabs on Society
Predictive policing systems combine sensor networks, big data and algorithms with the aim to predict and even prevent crime. But this type of pervasive surveillance raises concerns about privacy and profiling.

The ambition of the Dutch National Police (KLPD) is to realize a nation-wide sensor network that collects data on a vast scale both in the physical world and online. For its predictive policing ambition the KLPD received a nomination for the annual Big Brother Awards.

The aim of the sensor network is to greatly expand the police's observation capabilities. With the aid of algorithms to process the data, the network would be used for four types of observation:
  1. identifying people and objects;

  2. recognizing relationships between two or more people;

  3. recognizing unusual behavior and;

  4. intercepting communications.


Those are all procedures central to police work. The first three are probably practiced by police officers every day. However, today's limited resources force the police to focus on a small group of people: people suspected of having committed a crime. A nation-wide sensor network would remove that limitation and enable the police to observe everybody. Always.

Although such a network will have quite an impact on society, the KLPD did not see fit to share its ambition with the public. The vision described above was formulated in an internal document in 2011 [Dutch] but was not publicly known until it was leaked earlier this year.

Big Brother Awards


The Big Brother Awards (BBA) go to those who “have excelled in the violation of privacy”. BBA ceremonies are held in many countries including Germany, Finland and Austria. The Dutch edition is organized annually by Bits of Freedom, a digital rights organization. Competition for the BBA is murderous these days but planning an all-pervading snooping apparatus and being secret about it, gets you a spot with the final three.

The KLPD had to forgo the audience award, but it did receive the jury award. Jury member and journalist Bart de Koning said about predictive policing: “placing this method at the core of their practice, the police targets unusual behavior instead of criminal behavior. Either you walk in line or you fall victim to police scrutiny. If a civilian deviates from the norm he becomes suspect and becomes an at-risk citizen" De Koning based that remark on the work of criminologist Marc Schuilenburg.

De Koning then presented the Chief of Police of Amsterdam Pieter-Jaap Aalsberg with the BBA trophy. Aalsberg received a big applause from the audience for showing up to collect the prize. In his acceptance speech he said he was genuinely happy with the prize because he believed transparency and a public debate about predictive policing are necessary in a democracy.

“But”, he continued, “there are dilemma's. What if we can predict domestic abuse of children? Do we want to use the methodology for that? These are questions I struggle with as a policeman. I too am convinced that a society with maximal crime prevention is one that has a minimum of freedom. On the other hand, a society that does not take care of the most vulnerable of its members isn't a worthy democracy.”

Aalsberg recalled the film Minority Report in which a police officer is tasked with preventing murders by arresting killers before the deed is done. When the officer is predicted to become a murderer himzelf, he starts to doubt whether predictive policing is such a good idea. Aalsberg pointed out the police is still very far removed from a Minority Report level of predictive policing. At the moment the Amsterdam police is using data analyses to identify area's with an increased risk of burglaries. Sending out extra patrols to those areas has reduced break-ins with 20%, he said.

But the authors of the KLPD vision document do seem to have taken their cue from Minority Report. The authors state that the sensor network could be used to “predict occurrences in the near future” which would enable cops to “take action proactively”.

De Koning ended with the closing words that predictive policing is dependent on profiling. He recalled the case of Robert McDaniel from Chicago who was visited by a police commander to tell him he was being watched and better not commit any crimes. It turned out he was considered a risk and had been placed on a 'heat list', despite the fact he had committed no crime. Being a black man in a bad neighborhood an algorithm had spewed out his name. “It is really scary that someone who has done nothing wrong is labeled like that”, said De Koning. “You run the risk such harassment will push him over the edge because the label prevents him from getting a job.”

Image: Bart de Koning (left) and Pieter-Jaap Aalsberg
Loading comments...
related items