Paper short abstract

This paper examines the conception, applications and uses of "predictive justice" tools, that aim at harmonizing decision-making in the legal field. It is based on a survey (N = 17) conducted in France amongst both the producers and users of one of these tools.

Paper long abstract

“Predictive justice” tools (Cohen et al, 2020), which enable the anticipation of likely outcomes of a case based on its characteristics, have been at the center of numerous debates (Garapon, Lassègue, 2018). While these technologies could help standardize decisions across the country and/or within a given court (Chen, Spamann, 2016), such standardization conflicts with legal norms (Dumoulin, 2022) and values (particularly the concept of the uniqueness of each case). “Predictive justice” tools are therefore widely criticized and rarely used by judges, even when they have access to them (Brayne, Christin, 2021 ; Licoppe, Dumoulin, 2019).

In France, despite some pilot projects (Vergès, Vial, 2022), these tools are not used in the courts. However, they are used by lawyers, who employ them particularly in mediation processes, which help reach an agreement between the parties and thus avoid going before a judge. In this sense, the use of “predictive justice” by lawyers could indeed facilitate standardization of outcomes in cases where negotiation is possible.

This paper examines the nature of this harmonization. It shows that the results produced by “predictive justice” are based on a categorization (Bowker, Star, 2000) that is sometimes oversimplified, which imposes numerous limitations on the consideration of the specificities of individual situations. It also examines how lawyers adopt these tools, and establishes that lawyers do not simply accept the tools’ results at face value, meaning that the tools do not automatically produce the standardization they aim for.

Paper short abstract

The current paper models the actors involved in the deployment of Distributed Acoustic Sensing in the UK using Rasmussen's Risk Management Framework. Social network analysis is applied to examine system dynamics and identify key actors that shape the development and governance of DAS.

Paper long abstract

Distributed Acoustic Sensing (DAS) is an emerging technology that operationalises datafication by detecting environmental vibrations along the length of dark fibres (unused optical fibre cables), generating continuous streams of analysable data. DAS is already being used to monitor railway tracks and seismic activity, to detect leaks in pipelines, and more. The flexibility and resilience of DAS infrastructure are drawing interest towards its integration into smart cities. The operation of DAS in smart cities is not yet known, but will likely depend on partnerships between various actors ranging from industry to government.

The current paper uses actor maps to model the interdependencies among actors involved in the deployment of DAS in smart cities across the system hierarchy proposed by the Risk Management Framework (Rasmussen, 1997). We have created an actor map representation to examine the ‘layout of decision-makers, planners and actors’ involved within the DAS sociotechnical system in the UK. We then employed social network metrics to interrogate actor prominence and system-level dynamics. By identifying both top-down and bottom-up influences that shape the deployment of DAS, the paper demonstrates how accountability gaps may emerge within the datafied state. We also propose systemic recommendations to promote the development of future, justice-driven governance structures aligned with public interests in regulatory oversight.

Paper short abstract

This submission investigates the role of data visualisation for agonistic public participation in climate assemblies. By studying Gipuzkoa’s Citizens’ Assembly, it unpacks how the conflict between the promise of sharing power and established procedures was articulated around accountability charts.

Paper long abstract

This article examines how data visualisation operates at the interface between participatory processes and established governance systems. Bridging democratic innovation and transition studies, we argue that climate assemblies can be conceived as moderate-radical niche regimes that simultaneously depend on, reinforce, and challenge incumbent political structures. We integrate perspectives from critical data studies and design to explore how data visualisation can enable—or constrain—the transformative and agonistic role of democratic innovations in the transition of governance systems. Finally, we provide a situated empirical account of the theoretical frameworks through an in-depth investigation of the Gipuzkoako Herritarren Batzarra, a climate assembly in the Basque Country.

The study combines desk research, semi-structured interviews, and observation of the assembly process. Focusing on the accountability phase, generally highly contested, allowed us to unpack the reinforcive, evolutive, and transformative imaginaries that different actors simultaneously held. Data visualisation played a central role, as accountability reports relied on colour coded tables to track acceptance and implementation of recommendations. These visualisations enabled citizens to scrutinise progress and enact their role as “guardians” of the proposals. However, their inscription in a process with a narrow focus on policy implementation constrained their potential for deeper contestation disrupting established governance systems.

The paper argues that data visualisation could support the transformative power of democratic innovations by connecting assemblies with broader publics, and by enabling challenging the conditions of participation from within. However, without intentional design and adequate resources, dataviz risk framing tensions around governance as a background rather than a field of struggle.

Paper short abstract

This paper examines the "simulative turn" in which Big Data and AI reshape logics and practices of governing. It argues that these reconfigure political epistemology, shifting political practice grounded in representation and conflict toward regimes centered on modelling, prediction and preemption.

Paper long abstract

Over the past decade, public sector and political bodies have increasingly developed and deployed algorithmic, data-driven technologies to inform decision-making. These include sandboxes, testbeds and, more recently, digital twins as experimental governance infrastructures. Entangled with narratives of evidence-based policymaking, smart government, and experimentalist governance, their deployment raises questions about democracy and the political itself. Data analytics is often framed as better able to anticipate citizens’ needs and interests than citizens themselves. In this sense, such technologies reflect Jean Baudrillard’s notion of the vanishing of representation in the age of simulation and Donna Haraway’s claim that simulation replaces representation in the information society, signaling a potential erosion of political representation. This paper asks: To what extent do algorithmic, data-driven technologies in the public sector instantiate simulation as a pervasive form of governmentality? And what are the implications of this simulative turn for the very idea of political representation? Methodologically, the study combines close reading and Foucauldian critical discourse analysis, drawing on urban governance in the EU as an empirical case. It traces how simulation has emerged as a political technique to execute political agendas. As simulations become embedded in governance and decision-making, they reinforce a technocratic paradigm that seeks to datafy and quantify the world. The emphasis on predictive, data-driven evidence signals a shift from a political practice grounded in representation and conflict toward modelling and pre-emption. This transformation reconfigures political epistemology, positioning the assumption of predictability of political opinion forming as the unquestioned legitimacy horizon of political deci-sion-making.