Special session 9:
On the More-Than-Human Politics of Prediction: Epistemologies, Ideologies, and Ontologies of Modeling Behavioral Futures

This session calls for papers that examine the intersection of data-driven predictive technologies and the governance of human and non-human behaviors. This intersection materializes in a range of sectors, from lending risks in finance and real estate to preventative healthcare strategies to, most problematically, the prediction of criminal behavior based on a host of either personal or circumstantial data. The geographic nature of these processes is often overlooked, but spatial relations are crucial for understanding the stakes of data-driven behavioral prediction. Housing mortgages draw on and shape the racialized nature of neighborhoods and forces of gentrification, health risks are indelibly tied to resource access, environmental toxicities, and the conditions of contagion, and the forecasting of crime shapes urban planning decisions and the spatial distributions of over-policing. This session is interested in tracking how prediction, forecasting, and behavior modulation strategies produce effects in their own spheres but also, importantly, become unmoored from these spheres as the strategies of data science increasingly share spatial logics across models of forecasting. Early predictive policing borrowed techniques from the seismology of earthquakes (Jefferson 2020), for example, and more recent and sophisticated crime forecasting algorithms draw on the insights of ant biologists to understand the environmental conditions that shape collective or ‘swarm’ behaviors (Calhoun 2023). Across a range of empirical topics, we seek to ask: How does data-driven behavioral prediction blur the line between humans and non-humans, and how do predictive technologies convert behavioral futures into spatial objects of analysis in ways that go beyond traditional modes of categorizing human behavior?