In an age of connected and data-intensive intelligent machines, the role of science and engineering in society is changing. Not only does AI research and innovation intervene in our living environments, it shapes the interactions, relations and behaviours that make up social and cultural life. At CIM, we conduct interdisciplinary research to understand how different instantiations of AI, affect relations between data, science, and society, and how in predicting the future, they are likely to affect its realization.
The re-emergence of artificial intelligence, automation, and algorithmic reason into public discourse is a ‘revolution’ in the original sense, an intellectual ‘circling back’ to previous eras of the 1960s and 1990s in which proponents of neural-network metaphors attempted to reject the rigorous structuring of human cognition into logical formalism and rule-based framings. However, 21st-century intelligent systems are not only mathematical abstractions but also socio-technical arrangements deployed as part of our everyday interactions, transactions and encounters in the so-called “real-world.”
The increasingly sophisticated computational models embedded in online platforms, mobile apps, and living laboratories are not just used to predict the future and increase efficiency, they also have a variety of socio-material, semiotic and normative effects on everyday, professional and public life, as AI research and innovation is likely to give rise to new capacities of sociotechnical systems to transform society. For this reason, AI-inflected socio-technical arrangements must be studied and understood not just abstractly and formally but through in-depth sensory, ethnographic, and historical investigations.
Furthermore, the development of new mechanisms for testing, assessing and evaluating intelligent socio-technical systems – such as testbeds for smart city technologies, living labs and the construction of interfaces for ‘explainable’ AI – forms a key part of today’s AI revolution. At CIM, we use interdisciplinary approaches, such as computational sociology and the social studies of testing, to contribute to evaluating the performance of AI in society. To envision forms of machine intelligence in which social and technological players come together, rather than replace one another, we must invent interdisciplinary collaborations which leverage the potential of the "ensemble" to produce new understandings, and translations between otherwise-isolated fields.
- Scoping issues in autonomous vehicle research at the science/society interface, Research development Project, University of Warwick, 2016-2017.
- More than Human Smart Cities (Cagatay Turkay)
- Investigating Interactive Visualisation Techniques for AI Explainability (Cagatay Turkay, Co-I with R. Henkin)
Publications & other outputs
Marres, N. (2020). Co-existence or displacement: Do street trials of intelligent vehicles test society? British Journal of Sociology, Put to the Test: Critical Evaluations of Testing, Special Issue edited by N. Marres and D. Stark.
- Castelle M. (2018) Social theory for generative networks (and vice versa), blog post, September 2018
Image credit: Surfacing social aspects of driverless cars in the driver-in-the-loop simulator, WMG (Photo by L. Oliviera, 2016), Workshop, University of Warwick, December 2016