Predictive Policing


Two masked man break open a window and pursue to climb inside. Before they can set their feet inside the house a bright light blinds them and they hear the words being yelled ‘Freeze!’. A police offer prevents a burglary before the man can even enter the house. The scenario may suggest a coincidence of the officer being in the right place at the right time, in reality they were guided by a software program that predicted the crime.

Although predictions shouldn’t be taken as an absolute reality, the hypermodern technique to predict the time and location of future crimes using sophisticated computer models and algorithms is becoming a more widespread police system. The goal is to transform policing from a reactive process to a proactive process. An important side note is to view the technique as complementary to police knowledge, experience and judgment of possible relevant information (Greengard, 2012).

Key is to use a Data Warehouse encompassing data of criminal activities and linking relevant data to it (Willems and Doeleman, 2014) Although it is essential to keep the model simple, otherwise the model might wind up saying nothing definitive, over time other variables could be added in the model, such as epidemiology data, climate and weather, economic factors and even spatial geography and census data (Greengard 2012). To find additional patterns that the human eye would fail to see, data mining is being used. Crime will always have a significant stochastic or random component, however the non-random component that crime has is being exploited (Greengard, 2012).

For example the Amsterdam Police has implemented the system Criminal Anticipation System (CAS). High-impact crimes (burglary, robbery) are predicted by creating heat maps with territories that span 125 meter by 125 meter. Using only the top 3% of territories high risks are given warmer colors than lower probabilities with colder colors. On the basis of this information resources are deployed to timeframes and areas that seem to matter. 40% of all burglaries and 60% of all robberies are predicted by using only 3% of the territories in Amsterdam (Willems and Doeleman, 2014).

Heat

The opportunities of the system are obvious, however a possible threat to the system is criminals may anticipate on predictions and alter their behavior. A drop of criminal activity in one hot territory may be only temporal as in this cat-and-mouse game criminals will look for new spots to base their criminal intentions. Privacy issues are also a problem that may concern the system. Following individuals could touch upon privacy issues (Greengard, 2012). Furthermore these techniques may be used to justify greater data collection and more surveillance. Finally a future in which civilians are being corrected by the state before they do something suspicious is not one we root for (Reve, 2015).

Van het Reve, J. ‘De toekomstpolitie’, De volkskrant, 26-9-2015.

Greengard, S. ‘Policing the future’, Communications of the ACM, Vol. 55, no 3, march 2012.

Willems, D., Doeleman, R. ‘Predictive Policing-Wens of Werkelijkheid?’ Het tijdschrift voor de politie, jg 76, no 4, 5, 2014.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: