Vrije Universiteit Amsterdam New ways of working Anastasia Sergeeva Predictive policing in the Dutch police force Line Nielsen 2680709 Suad Osman 2516952 Naomi Tilon 1722433 Carla Kabamba 1530623 Josefine Due Olsen 2680708 December 2019 1 Introduction How safe would it be if the police knew everything in advance and could prevent all forms of crime? Society would look very different. Making predictions has been there for more than a hundred years. In ancient Greece, predictions were made based on the signs of the gods. These days there is still a need for predictions. Making predictions has however changed into rational science as part of intelligent working. Technology is increasingly playing a role in the police force. The increasing amount of information that is available makes it possible to discover and explain patterns and trends on which we can make predictions. In 2013, the Crime Anticipation System (CAS) was introduced to the Dutch police force, and is currently being used in almost 168 Dutch police stations (Waardenburg, Sergeeva & Huysman, 2018b). CAS predicts where and when crimes such as street robbery and burglary will take place. As a result, the police works in a more targeted way, which can lead to less crime. Working in a more targeted way has even become more crucial over the years, as the Dutch police force have been struggling with a staff shortage for some time (NOS, 2019). In this assignment, we wish to examine how CAS and predictive policing has changed the work within the police force. We found that the introduction of CAS especially changed the profession of the information officers, who are now titled as intelligence officers. This lead to the following research question: “How has predictive policing affected intelligence officers in the Dutch police force?” Since this paper focuses merely on the occupation of an intelligence officer, Rudi Volti’s theory from 2011 is used to analyze the collected empirical papers, which are introduced later in the paper. The assignment is limited to answering the research question, knowing that the discussion on predictive policing is far greater. 2 Background Digital technology in crime prevention In the traditional police work, officers would patrol on the streets looking for crimes that would occur or that were occurring. The focus in the 60s was on random surveillance and fast response times. Officers were mostly dependent on their gut feelings, observations, and information they received from the public on a crime that would be committed. The work was reactive, in terms of officers responding to a variety of situations as they develop. In the 1980s there was a shift towards community led policing with the focus to be problem solving and to achieve proximity. Later in the mid-90s, for the first time, crime was viewed geographically through hotspot maps. With this new form of policing, they tried to better target the police by visualizing the crime intensity in geographical areas (Rinks, 2015). Nowadays, officers are even more dependent on effective information gathering. Law enforcement agencies have adopted new technologies in order to control crime. Predictive policing refers to the usage of analytical techniques to identify places and times with a high crime risk (RAND Corporation, 2013). The more data the system has, the more accurate predictions can be. In addition to the number of incidents from the past, the system also takes into account the change in the number of incidents at that location over time and other factors such as the presence of a police station. With the usage of predictive policing law enforcement tries to identify the pattern of crimes to be able to take pre-emptive action. In addition, the transfer towards predictive policing was followed by the emergence of the “intelligence officer”. Intelligence officers were supposed to support police officers in the use of predictive policing technology, by helping them to make sense of algorithmic outputs. Technology continues to change the profession of law enforcement. The day-to-day patrolling is effectively allocated to districts and neighbourhoods based on predictive policing, where and when crime is most likely to occur. The goal of Criminal Anticipation System (CAS) within the Dutch police force is to provide relevant and useful information to optimize patrolling on the street. The software lays a grid over a city with squares of 125 by 125 meters and then determines the risk of crime for each square. Where there is a greater chance of crime, the box turns darker. Colleagues are advised, to pay more attention to certain areas than to others because the calculated chance that incidents will occur there is simply higher (Doeleman & Willems, 2014). 3 Method & Theoretical framework The Trait Approach The research on which this report is based focuses on professions and occupational change due to technology. Certain kinds of workers belong to a particular occupational category that allows being termed by “professional”. What made profession distinguish from other occupational categories is discussed by sociologists through time. A number of sociologists and other researchers have created this so called “trait” or “attributes” approach, which is a kind of checklist to determine if an occupation can be characterized as a profession. In relation to this approach, there is a general agreement on the key traits of a profession between sociologists and other researchers. At first, a profession requires specialized knowledge, which relies on a wellestablished theoretical foundation. Acquiring this knowledge takes time, effort, and formal guidelines, which is why the second key trait of a profession is a university-based training that makes it possible for individuals to become professional practitioners. Generally, this means an education in a university setting. Third, the work of professional has to be of great value of the services to both society as a whole and as an individual. As the fourth key feature of profession is code of ethics, which consider the importance that the professionals do not use power for the wrong purposes, but instead they put the needs of society and individuals ahead of their own interests. Fifth, relationships among professionals are expected to be cooperative and, in some cases, competitive, but their actions must be governed by commonly accepted ethical standards of professionals. There are many more features discussed by researchers, but as the last common key trait is autonomy and self-governance. The professionals’ ability to act with a high degree of autonomy and governance, which relate to the authority of professionals and being accepted into the ranks of a special profession (Volti, 2011). The Professional Continuum Later on, the early approach was criticized for only being appropriated for established professions of medicine and law. While taken other occupations into consideration, the lines between occupations and professions were blurred – some of them only fulfilled several traits and others all of them but in attenuated form. For this reason, the terms of semi-professions and para-professions were established and commonly used. Semi-professions are characterized by only fulfilling some of the attributes or to a weak extent (Etzioni, 1987). Para-professions refer to occupations that have some of the traits but mostly are assistants to an established profession. The work of paraprofessionals involves routine activities, which are essential for the profession to be effective. Important to point out is that the boundaries between profession, semi-profession and para-profession are indistinct (Volti, 2011). However, earlier researchers failed to give sufficient attention to the political processes of professionalization in which some occupations were able to achieve professional status. Certain occupations and their members gained the prestige, autonomy, and other advantages to attain professional status. The political process leads to the power dimension of professions. The most salient of a profession is the ability to control the most important aspects of the work they do. This ability concerns to determine following; who is allowed to do the work, how they are prepared to do the work, and who gets to evaluate their performance. Furthermore, professionalization as a means of control also relate to the control of entry into a profession in which the role of education and licensing are essential (Volti, 2011). Due to technology and how it may change occupations and professions, it can basically be analyzed through two views; can be a threat and can be leveraged. Technology challenges the profession or power in control in ways of redistributing tasks to clients, performing tasks automatically, or allowing semi-professions to do complex work. Besides, new technologies also have reinforced the status and authority of professionals and made it possible to deepen the specialization. Research setting A report on an ongoing ethnographic study at the Dutch police is collected to support our research question (Waardenburg, Sergeeva & Huysman, 2018b). The study focuses on the activities of the so-called “Criminal Anticipation System” (CAS), which is a system of algorithms to identify patterns in existing data and creates a heat map that shows the most vulnerable places to crime. Findings of the study indicate that the shift from traditional police work to predictive policing has affected an emergence of a novel occupational role “Intelligence officers”. Concerning occupational change due to technology, previous research access two views; an occupation transforming the expertise and tensions or conflicts between occupations. Thus, the study emphasizes what happened when introducing the new occupational role of intelligence officers to support operational police officers in the shift from daily police practices to data-driven work. Another study of Waardenburg, Sergeeva, and Huysman (2018a) assesses that the introduction of predictive policing requires police officers to alter their focus to more data-driven work instead of the traditional way by looking at criminals’ social environment. To understand and predict crime, complex analyses and algorithms are applied, and because of that a high amount of certain knowledge is required. The study is relevant for our research to emphasize the importance of support from intelligence officers in the altering to more data-driven work. 4 Analysis and results To illustrate how predictive policing has affected the work of intelligence officers, it’s relevant to look at how general police work has changed. As mentioned earlier, intelligence officers gradually have more influence on the daily work of a police officer. From Volti’s checklist approach to profession, traditional police work is defined as a profession (Volti, 2011). They have specialized knowledge which they gain from a specific education program, the service that they provide is of value to society and is not broken due to their code of ethics. They have autonomy and authority over the public. The intelligence worker however, has no need to have specialized knowledge within the police force, since the same algorithms are used in marketing (Waardenburg, L., Sergeeva, A., & Huysman, M. 2018b). In fact, intelligence officers have no requirement to be trained police officers. Furthermore, they have no authority over clients, since they have no real interactions with the public. While distinct in terms of degree of professionalism, due to technology, the two types of work blur more and more together, herby challenging the traditional police work. Before CAS was introduced, there was a clear line between the police officers and the information officers in terms of power. Now, because of the need for analytical, meaningful work, these roles have been redefined by shifting a lot of power and responsibility to these ‘semi-professionals’, the intelligence officers. Semi-professionals are now able to do specialized work, without having the specialized background for it. This does not only change the daily routine for occupational, it also challenges the status and authority that has been prevalent within the police force up until now. In addition to the professional continuum, a shift also occurred in terms of hierarchy, which in turn challenged many of the traditional roles norms. Due to aforementioned technological developments in predictive policing, of which, in this case, CAS is mostly focused on, the occupational tasks of information officers changed from a position in the back-office department to a more valued role in the office. In the introductory stages of predictive policing methods these information officers were believed to be able to bridge the gap between ‘data scientists and police officers’ (Waardenburg, L., Sergeeva, A., & Huysman, M. 2018b, p. 7). As the shift from a reactionary approach to a more predictive one occurred, by being able to read, interpret and integrate algorithmic outputs onto police work, the role of the information officer became much more valued and important. In managing these databased system the information officer stood more at the forefront in predictive policing and consequently a different order in hierarchy occurred. Relating to the two views mentioned earlier, this technological development could be seen as an advantage to the information officer as the technology deepens his specialisation and with it, may have a higher and more prestigious role within the police force. Consequently, the new role of the traditional police officer consisted of a set of tasks that were described as ‘a messy and random list of activities’ (Waardenburg, L., Sergeeva, A., & Huysman, M. 2018b, p. 11). This shows that, besides experiencing difficulty in integration of the CAS system onto their work, the traditional police officer found himself somewhat driftlessly manoeuvring through these innovations. Although, future police officers are somewhat more prepared to a technologically enhanced future, through the use of ‘advanced audio-visual technologies to diffuse professorial wisdom’ the question of how real-life experiments aided by CAS are to be fully integrated in the police force still exists (Volti, 2011, p.11). In other words the traditional police officer experienced a decrease in prestige and autonomy due to these changes. Lastly, these innovations caused data-driven decisions to be greatly based on the collective findings of police officers and the individual judgement of a police officer to be quickly found inferior. As a police officer generally becomes much more dependent on these systems, the officer views it to be ‘superior and of higher purpose’ than its own tacit system of knowledge as the effects and impacts of these systems are found to be much more precise and efficient compared to former ways. In disregarding the set of traditional police tasks and the tacit knowledge that is seen as an important element of police work, these innovative systems are centralized even more so by the traditional police officer. Emergence of intelligence police officers The transfer towards predictive policing was followed by the emergence of the “intelligence officer”. Initially, intelligence officers were supporting police officers in the use of predictive policing technology, by helping them to interpret the algorithmic outputs. However, by investing a lot of expertise into interpreting and translating algorithms and the outputs, intelligence officers became increasingly influential and started to steer police action (Waardenburg, Sergeeva & Huysman, 2018b). The predictions based on these analytic methods guide the decision-making of law enforcement agencies, especially with allocating resources to areas that show an increased crime risk. With the emergence of the intelligence police officer, the police force aims to move policing from reactionary investigation to proactive investigation (National Crime Prevention Council, 2006). As mentioned, Volti (2011) argues that the introduction of algorithmic analytical tools, can either be a threat or leveraged to a profession. It can be argued, that the introduction of CAS is viewed as leveraged to both the operational and intelligence officers. Technology is not taking over the actual occupation, thereby not being an actual threat to the professions. However, as mentioned earlier, the introduction of CAS lead to a huge shift within the sphere of hierarchical norms, thereby pressuring the operational officers to change their knowledge while allowing semi-professions to have more authority over their decision-making (Volti, 2011). 5 Recommendations Emphasis on predictive aspect of policing Predictive policing through the CAS system mainly depends on knowledge and data sources to which it is exposed. This implies that the use of these models are somewhat only functional on the premise of past events. These systems, therefore, rarely rely on contextual changes in local crime. As mentioned in Degelin’s article on the effectiveness of predictive policing: ‘burglars looked for new areas and did not behave as expected, though they were clearly serial offenders’. In this sense, predictive policing has an indirect effect on the behaviour of criminal offenders. Although studies have proven forms of biases exist within the predictive policing systems, there has not been enough emphasis on the prevention of these types of behaviours, and of other external influences unrelated to new technological approaches. Given that the occurrence of crime does not mainly depend on past events, a new way of technologically enhanced approach on policing may be necessary. Advice regarding this aspect would be to include traditional tasks of the police officers, but to develop existing digital technologies that may be able to aid and reinforce traditional methods of policing. Specialized educational framework Considering the fact that the lines between intelligence officers and police officers are blurred, but the intelligence officers do not need university based-training, a specialized educational framework is a logical next step for this relatively new function. It would legitimize the position of intelligence officers more as equals of police officers and this equalization can take away existing tensions between both groups, thus bridging the gap. Code of ethics Intelligence officers are finding their way into police work in an organic but unregulated way, which can cause the quality of their work and ways of working to differ from place to place. They also deal with sensitive data and need to be held accountable just like police officers, to help ensure an agreed upon level of quality in a city or even country. Autonomy and authority over data With the obligation of being held accountable, autonomy and authority over the data is a more viable option for intelligence officers. While being held to a higher standard, supervised, and gaining influence because of their knowledge regarding data, a fitting increase in responsibility seems to be a natural next step. Independent supervision Police officers are supervised, this is common knowledge, and this supervision ensures a level of quality and adherence to the law. Intelligence officers deal with sensitive data and influence how law enforcement carries out their duties, so independent supervision will add necessary levels of safety to the data. In addition, it levels the playing field for police officers who can only do their job with supervision. Further training The police officers are still getting used to the increased influence of intelligence officers, there is knowledge gap between what the intelligence officers do, and what the police officers know about it. This brings us to further training. Further training can bridge the gap between traditional policing and the current databased work methods. This prevents the two groups from alienating each other and can address any tensions based on lack of knowledge or understanding about each other’s fields of work. Reorganisation In order to safeguard the tacit knowledge of police officers, while allowing the space to improve technology in this field, a reorganization can offer solutions. Though, police officers are adjusting to and need to learn about databased work, their tacit knowledge is not obsolete and can have a place in an innovating police force. 6 References Cope, N. (2004). Intelligence Led Policing or Policing Led Intelligence?: Integrating Volume Crime Analysis into Policing. British Journal of Criminology, 44(2), 188–203. doi: 10.1093/bjc/44.2.188 Doeleman, R., Willems, D. (2014). Predictive Policing – wens of werkelijkheid? Het Tijdschrift voor de Politie. NOS (2019). Politiebonden doen dringende oproep: 'Maak keuze bij inzet agenten'. Retrieved November 27, 2019, from https://nos.nl/artikel/2311389-politiebonden-doen-dringende-oproepmaak-keuze-bij-inzet-agenten.html RAND corporation. (2013). Predictive Policing: Forecasting Crime for Law Enforcement. Rinks, R. (2015). Predictive Policing: Kansen voor een veiligere toekomst. Lectoraat Intelligence. Volti, R. (2011). Professions and professionalization. Chapter in: “An introduction to the sociology of work and occupations” Sage Publications. Waardenburg, L., Sergeeva, A., & Huysman, M. (2018a). Digitizing crime: How the use of predictive policing influences police work practices. Paper presented at 34th European Group for Organizational Studies (EGOS) Colloquium, Tallinn, Estonia. Waardenburg, L., Sergeeva, A., & Huysman, M. (2018b). Hotspots and Blind Spots: A Case of Predictive Policing in Practice. Springer New York LLC. 7 Appendix We chose predictive policing as research topic to gain more knowledge on how technology has changed the police work over time. What are the differences between the traditional and new ways of preventing crime within the police force? Google scholar was used to find relevant articles. Our search terms were for example: “predictive policing”, “forecasting crime”, “police information technology”, “police work transforming”, “impact of technology on policing”, “intelligence led policing”. The focus in our paper is the Dutch police force as the Netherlands uses ‘predictive policing’ on a large scale. Furthermore, for the past few years Dutch police force is under staff shortage. The papers of Doeleman and Willems(2014), and Rinks(2015) provide insights on the implementation of predictive policing in the Dutch police force. Crime Anticipation System (CAS) is being introduced further across the Netherlands and more than 160 teams are going to work with the system. The papers of Waardenburg, Sergeeva & Huysman (2018a, 2018b) discuss predictive policing in practice, in particular CAS within the Dutch police force. We found that the introduction of CAS especially changed the profession of the intelligence officers. This led to the research question: how has predictive policing affected intelligence officers in the Dutch police force? Our paper focuses on the digital transformation of this new profession and what traits the intelligence officer as a profession has, based on the checklist of Volti (2011). Furthermore, the introduction of predictive policing changed the police work into more data-driven work instead of the traditional way of relying on the public to provide information.