How can we secure an accessible and open democratic debate about police use of predictive analytics when the technology itself is a specialized area of expertise? Police utilize technologies of prediction and automation where the underlying technology is often a machine learning (ML) model. The article argues that important issues concerning ML decision models can be unveiled without detailed knowledge about the learning algorithm, empowering non-ML experts and stakeholders in debates over if, and how to, include them, for example, in the form of predictive policing. Non-ML experts can, and should, review ML models. We provide a ‘toolbox’ of questions about three elements of a decision model that can be fruitfully scrutinized by non-ML experts: the learning data, the learning goal, and constructivism. Showing this room for fruitful criticism can empower non-ML experts and improve democratic accountability when using ML models in policing.

implies that scientific publications are made freely accessible on the web. The author or originator keeps the copyright to the publication, but gives the users permission to read, download, copy, distribute, print out, search or link to the full text without a claim for compensation.

Inter-organisational and inter-disciplinary investigations are increasingly deployed against criminal networks and cross-jurisdictional crimes. This study provides a unique empirical window into an inter-organisational investigation against a large organized crime network in Norway. Building on interview data from the participants in the multiagency investigation team that was summoned for this case, the article discusses coordination issues that arise when organisations with different goals and interests collaborate to reach a common goal. The article studies co-ordination from inside of the investigation team and discusses the interchangeable use of criminal and administrative law. While bridging organisational boundaries enable agencies to pool powers, coordination across organisations may challenge the protection of sometimes conflicting aims and interests.Keywords: organisational boundaries, inter-organisational co-ordination, security networks, policing 2 IntroductionOslo area, September 2014: After a period of comprehensive communication surveillance and covert investigation, 280 officers from the police, the Tax Administration (henceforth 'TA'), and the Labour and Welfare Administration ('Labour & Welfare') carried out a raid on 20 shops in the 'Lime' grocery chain. The grocery chain was established and operated by an organised criminal network that was previously known to the police. The shops were profitable to the network who allegedly utilised human trafficking for forced labour, illicit work, money laundering, benefit fraud, tax evasion, violations of immigration law, employment offences, identity theft, and credit card fraud i . The raid marked the beginning of a prolonged multi-agency investigation, covering a variety of criminal and administrative violations. The investigation was structured as a multi-agency project, which meant that investigators from all participating agencies (police, TA, Labour & Welfare) were removed from their daily tasks, co-localised in a shared, rented office space, and co-ordinated by the Lime project's chief investigator.Pluralisation characterises contemporary policing, and partnerships as well as networked security governance between state and non-state actors have been examined empirically (Gundhus et al., 2008, Nøkleberg, 2016, Søgaard et al., 2016. This study explores the co-ordination of state powers in the multi-agency Lime project. While the literature on the pluralisation of policing and nodal governance of security (Johnston andWood, 2003) emphasises the role of non-state agents in the provision of security, this study delves into the heterogeneity of powers and interests in state sector governmental nodes. Partnership working and information sharing among public agencies intensifies 'the state gaze' […] in a way which is valued for its ability to make the daily work of the various state agents easier. '(O'Neill and Loftus, 2013, p. 451). This study contributes to the existing literature in two ways: Empirically, through studying the internal 3 operations of...

Drawing on narrative criminology and sensemaking theory, this paper explores interpretive patterns in an interagency policing collaboration that targets ‘work-related crime’ (WRC). WRC is a policy term denoting organized crime and economic offences (i.e. tax evasion, benefits fraud, labour exploitation and immigration law offences) and is framed as a threat to the viability of the welfare state. While the concept signals an intent to coordinate across agencies, policing takes place within local and institutional contexts. How do agents in the police and collaborating agencies render the policy concept ‘work-related crime’ meaningful and actionable? The study articulates three organizational narratives explaining WRC as fundamental criminogenic change, as stability and as a reflexive product of the control apparatus.