Coexistence: Evidence-Based Policing and Emerging Technologies
Valarie Findlay – Research Fellow, Police Foundation (US)
Evidence-based policing promises to be a powerful force for change and holding significant value for police organizations, the public and community safety. Often misunderstood, evidenced-based policing promotes the importance of practices based on scientific evidence that demonstrates what works ‘best’.
Plainly stated (sort of), qualitative and quantitative data is gathered from operational practices in a controlled framework that accounts for fluid environmental factors. Later, this data is analysed to fact-based evidence used to improve policies and procedures.
What evidence-based policing is not is the uncontrolled, empirical collection of data or its arbitrary use to implement or change procedural approaches and practices. Properly constructed, it does the opposite.
Allowing the integration of other research methods, such as discourse analysis, can create context where dynamic factors influence responses, such as offender interactions that may differ depending on circumstance, individual, neighborhood or city.
A linkage between research and practice, evidence-based practices hinge on the value of anecdotal and formalised evidence from measurable policing functions – from implementing guidelines to crime prevention. This nexus of controlled theory married with practice outcomes is gaining traction as it allows for faster operationalisation of academic research. From this, the best quality evidence can be used to shape ‘best practices’ through a process of re-testing and measuring the hypothesis.
Looking at both static and dynamic factors, the ones that change analogue and expected outcomes, is valuable to policing since its a human-behaviour driven profession. It’s easy to presume that under a particular scenario that under “normal” circumstances any person would react a certain way or that from the officer’s “experience” a certain outcome is presumed; this is what evidence-based policing stands to clarify.
The field of medicine is a good example. It’s widely accepted by professionals that diagnostics rely on three important areas of ‘evidence’: clinical signs, pathology and anecdotal (patient-view) since how disease presents itself physically, biologically and practically varies by individual. These diagnostic axioms produce irrefutable facts that remain relevant despite the use of technology; noting, centuries ago diagnostics were largely anecdotal and treatment was empirical.
For policing, an evidenced-based approach could look like this: a hypothesis is drawn from anecdotal evidence showing community-based policing programs deter recruitment of youth in street-level gangs. A sample group or area is selected to test the hypothesis by collecting data from practice activities by the subject matter experts (officers). Controlled by research methods that isolate criteria (demographic, circumstantial, geographic, etc) findings can be re-tested with outcomes confirmed, creating procedural or programmatic approaches.
Turning to technologies and how they could assist in evidence-based approaches, I’ll preface this by saying identify the “problem” and organisational requirements and letting those drive the selection of the technology; opposite view of how technology can be used to address all problems only serves the vendors.
Using artificial intelligence (AI) as an example – a hot topic across all sectors – may solve one of policing’s longstanding problems with the persistence of analogue data alongside electronic data and converting it into usable intelligence. AI could play a key role in addressing this problem not necessarily as an approach that automates decisions, actions and responses, but aiding in them.
Clearly, the more integrated public safety data is, whether individually identifying or statistically informing of a group or area, the more informative it is. AI can create generative processes that correlate large amounts of data, replicating ‘human’ thinking through algorithms and digitized heuristics, learning from outcomes. These predictive, adaptive responses can then be extended to similar problems making AI inherently useful to some aspects of policing.
The value to evidenced-based practices could improve data analytics across many programs – crime prevention and interdiction, investigative functions in analysis and intelligence gathering, from tip consolidation to complex field intelligence.
Despite AI’s ability to rapidly correlate and create associations with data, as with any technology, it has limitations and risks. The outputs and conclusions generated by AI processes are only as good as the input data and algorithms – it is truly a matter of ‘garbage in, garbage out’ and AI will not magically improve the quality of input data.
While ethics and privacy issues arise with AI’s ambiguous processes, for evidenced-based policing to explore its value, high-integrity algorithm, validation and verification will be key to establishing trust and confidence for all sides.
Valarie Findlay is an ASEBP Member and a research fellow for the Police Foundation (USA) with two decades of senior expertise in cyber security and policing initiatives. She has worked extensively on federal cyber initiatives and is a member of the Canadian Association of Chiefs of Police eCrimes Cyber Council, as well as other federal cyber councils in Canada and the US. She holds a Masters in Sociology and a Masters in Terrorism Studies from the University of St. Andrews; her dissertation, “The Impact of Terrorism on the Transformation of Law Enforcement” examined the transformation of law enforcement in Western Nations.