250 words agree or disagree
While I have mixed opinions about predictive policing, I find it to be an incredibly fascinating topic. Personally, I think it can blur the lines of civil liberties protection and proactive policing. However, I think with the proper, no-kidding, honest-to-goodness checks and balances, the technologies available to law enforcement and intelligence agencies can indeed be a force multiplier and allow for effective prevention of criminal acts or terrorism.
The concept of protective policing brings to mind a couple of things. First and foremost, that law enforcement agencies utilize data science to create algorithms that conduct trend analysis and/or heat maps to indicate areas of significant instances of criminal activity. For example, in my town, statistically more crime occurs in low-income areas and housing projects. Our police department uses criminal intelligence analysts to do everything from social media/open-source analysis to trend analysis to perform predictive policing. One benefit of watching social media is that over time, criminal tradecraft can be identified, such as drug deals, communications methods between persons of interest, and neighbors reporting activity (to name a few). Online databases and scripting tools allow data scientists to create algorithms to identify, display, and alert users to endlessly-customizable circumstances. For example, when a known subject (SUBJ) is mentioned or tagged on social media, or when he checks in at a certain location.
Another example is forensic trend analysis. One downside of this approach is that it, by default, requires historic analysis. This, then, requires that before such analysis is performed, it requires the long-term collection of information on activities… which means that efficacy of “preventive” policing evolves from a baseline to more effectiveness over time as more collection and analysis is performed. What concerns me about this, as I touched on before, is the potential for abuse. One “lesson learned” from an overseas tour was that often, people with differences of any sort (whether it be from perceived/actual personal slights, racial differences, tribal conflicts, etc), often report derogatory information about their neighbors for some form of personal gain – satisfaction, monetary compensation, or some other motivation. This, then, results in that neighbor being arrested, interrogated, or detained for long periods of time simply because a neighbor had a proverbial axe to grind. This is an inherent flaw in proposed “red flag laws” which come from a good place of course, but in practice, humans are fallible, and implementing a system where such “poison pen attacks” don’t result in an innocent person’s civil liberties being infringed upon.
Geospatial Predictive Analysis is an interesting new trend that I can get behind. There aren’t any civil liberties to be concerned about or trampled on, and can all be done remotely, albeit after-the-fact or at best, in near-real-time. Lawful intercept operations can help here as well. Incidents can be plotted on a heatmap, included in a .kml (google earth geographic mapping file format), and analyzed in various ways and correlated with various data to not only identify who was where and when, but also what electronic signatures were present at the time of the occurrence. These can be flagged or marked so that when these correlations once again align, police can be dispatched to a given location (say, an area with known drug activity or gang violence).
These same tools can be used to identify indicators not necessarily associated with individuals, such as time, locations, and other factors that can be identified and then weighed to determine whether things such as poor lighting, income of a given neighborhood, time of day, etc., are able to be identified as pre-event indicators that police can then use to attempt to mitigate risk. This might include visiting a neighborhood and encouraging a populace to take mitigation methods such as locking or alarm devices, or lighting improvement. A less-involved approach may be as simple as increased patrols when or where indicators are present.
Beck, C., & McCue, C. (2009). Predictive policing: what can we learn from Wal-Mart and Amazon about fighting crime in a recession?. Police Chief, 76(11), 18.
Department of Homeland Security. (n.d.). Dhs.gov. Retrieved from Fusion Center Success Stories: www.dhs.gov/fusion-center-success-stories
McCue, C., PhD., Miller, L., & Lambert, S. (2015, September 08). The Northern Virginia Military Shooting Series: Operational Validation of Geospatial Predictive Analytics. Retrieved August 6, 2019, from info/northern-virginia-military-shooting-series-operational-validation-geospatial-predictive-analytics” target=”_blank”>https://preventviolentextremism.info/northern-virginia-military-shooting-series-operational-validation-geospatial-predictive-analytics