... To what extent do the means of computation, from statistical models to learning algorithms employed in predictive analytics change this relationship, and the collective capacity and legitimacy to engage with future, uncertain situations? How do technologies of prediction change policies? Who predicts, how, and with what effects on decisions and administration and on their politics? More generally, how do ways of predicting institutionalize, fail to or change?
The ‘Governing by prediction?’ conference brought together several dozens of social scientists from around Europe and the US in September 2017, to investigate the ways in which data is produced, processed and modelled as part of the governance of collective problems - from chemicals risks to terrorism, in passing by climate warming and control of epidemics. This international event concluded the ANR-funded INNOX research project. On the agenda of participants were the following questions: What is the characteristic of today’s relationship with futures in policy making and action? ...

1st keynote address
Steve Hilgartner // Ways of predicting and limiting the consequences of hurricanes

On how physical simulations of hurricanes gain credibility, circulate, and perform the governance and management of extreme events

Session 1 // Communities and Agendas of Modelling for Policy

With Pierre-Benoit Joly and Bruno Dorin, LISIS and CIRED, on the political economy of global agricultural modelling
Christophe Cassen, Béatrice Cointe and Alain Nadai, CIRED, on organizing policy relevant expertise for climate, and the community of Integrating Assessment Modelling
Sam Randalls, University College London, on the predictive politics of the modelling and generation of resilience
Stefan Aykut, LISIS, on scenarios as infrastructures of policy change in energy systems, in France and Germany

2nd keynote address
Claudia Aradau // Governing others: Anomaly and the algorithmic subject of security

On how digital technologies and algorithmic rationalities reconfigure security practices, and knowledge of the "other".

Session 2 // Tools of computation and prediction, and their politics

With Bilel Benbouzid, LISIS, on predictive policing : science, organization and law
David Guéranger, LATTS, on technicians, elected officials and software in regional road network policy
François Dedieu and Sylvain Parasie, LISIS, on conflicts and complementarities between citizen data and regulatory air quality monitoring

Session 3 // Machine learning as Prediction

Dominique Cardon, Sciences Po Medialab, on personalized prediction and machine learning methods in tools for web computation
Adrian Mackenzie, Lancaster University, on techniques of prediction and generalization of complex populations

Session 4 // Counting, monitoring and predicting risks

Francis Lee, University of Uppsala, on the ambiguities of algorithms, data and judgment in epidemiology and disease monitoring
Henri Boullier on the economy of predictive knowledge, and software tools to predict chemicals risk in REACH

Session 5 // Simulation and Anticipation

With Céline Granjou, IRSTEA Grenoble, on ecotrons as infrastructures of environmental anticipation
Grégoire Mallard, Graduate Institute, on the art of simulation in disarmament talks
Dirk Scheer, Karlsruhe Institute of Technology, on conceptualizing the relation between scientific simulation and policy-making

Closing keynote address
Paul Edwards // Knowledge Infrastructures under Siege: Environmental Data Systems
as Memory, Truce, and Target

From critiques of simulation to climate change denial: how environmental data systems have become political targets