In the current Canadian carceral system, categories of risk related to reoffending are assigned to individuals based on their behaviours and traits inside and outside carceral institutions. This dissertation examines the complex history and application of how classifications are generated to understand how the police take up these classifications in their work toward ensuring public safety as they supervise certain high risk individuals in the community. An 'action in practice' methodology (actor network theory) was adopted in this project whereby voluminous government-produced records were analyzed and interviews were used to capture the experiences of police officers in Edmonton, Alberta, Canada, and how their work relates to risk classification practices for the supervision of individuals classified as 'high risk to reoffend' in the community. This research captures how police officers self-identify and define their roles in risk classification practices through which both actions of replication of carceral risk classifications and altered policing practices have emerged. This original research makes visible how police officers' processes and practices exist within the parameters of the current carceral risk classification system. By challenging positivist, realist notions that risk is a negative and calculable construct, this research instead engages with constructivist risk theory, drawing specifically from Ulrich Beck's risk society theory. The resulting analysis utilizes narratives, stories, and experiences of risk as defined by the police in their writings and conversations to construct a more comprehensive understanding of risk classification practices at the level of municipal policing. This dissertation is organized into four core chapters addressing: (1) risk is essentially a psychological-statistical construct premised on individual traits or behaviours or combinations thereof; (2) risk classification practices are legitimized by a system-wide deference to those who claim to possess expert or scientific knowledge on risk; and (3) the lived realities of high risk individuals, the police, and Canadian communities more generally spotlight the inapplicability of risk classifications in alternate contexts outside of carceral institutions. The findings from this research contribute new sociological insights into the reality of altered practices and innovations in policing today through a focused case study of a municipal police force in Canada.