Predictive Policing: Brave New World

In one of our previous posts, we discussed biometric technology and the role it plays in Canadian law enforcement. It is, however, only one of the “predictive” tools utilized by the police in relation to criminal investigations.

A new report by the Citizen Lab at the University of Toronto goes into alarming detail regarding growth of algorithmic policing methods, and how this technology compromises the privacy rights of Canadian citizens. The report is incredibly thorough and comprehensive, delving into how this controversial technique offends various sections of our Canadian Charter of Rights and Freedoms. Firstly, though, it is important that our readers understand what algorithmic policing is.

The overall success of any algorithm is the system’s ability to gather, store, and analyze data – with law enforcement’s methodology being no different.  A “location focused” algorithmic approach seeks to determine (predict) which areas are more likely to see criminal activity. The algorithmic system in these pursuits analyzes historical police data to identify geographical locations where crimes are, in theory, more likely to be committed. If this sounds familiar to you, then you’ve likely heard of, or accessed, the Vancouver Police Department’s GeoDash crime map – an online tool where you can navigate a map of the City of Vancouver by crime occurrence. You can choose from a variety of offences on the dropdown list, including homicide, break and enter, mischief, theft, and “offences against a person” which likely includes a variety of crimes such as sexual assault, assault causing bodily harm, and uttering threats. By looking at this map, you get an idea of which neighborhoods in Vancouver are most vulnerable to crime – except that it’s a little bit more sophisticated than that, and goes far beyond simply dropping a pin on the map. The public can see where the crime took place, but not who is alleged to have committed it.  The offender’s personal information is logged, in as much detail as possible, and becomes part of a larger system dedicated to predictive surveillance – i.e., it creates a profile of which individuals are more likely to commit a particular crime. This profile can be used to identify people who are “more likely to be involved in potential criminal activity, or to assess an identified person for their purported risk of engaging in criminal activity in the future”.

While this information is definitely concerning, there is another issue:  we have very little insight into the extent that this technology is being used. We know that the methods by which police gather information have historically discriminated against minority groups and those living in marginalized communities. This seems to guarantee that the VPD’s use of algorithmic investigative techniques relies on data that is often obtained through biased methods. We know that black and indigenous individuals are disproportionately represented in the correctional system, which can only mean that they are disproportionately represented in respect of these algorithms.

Although not everyone agrees that systemic racism exists within the VPD, the calls to address, unravel and mitigate the harm to marginalized groups continue to amplify. The idea that information collected under the apprehension of bias will not only remain on record, but will be used to further future investigations, is an indicator that Canadian law enforcement’s road to redemption will likely be a bumpy one.