The Targeting State: AI, Surveillance, and Predictive Power
Across migration enforcement regimes and contemporary security infrastructures in the Global North, AI-driven systems are becoming deeply embedded in the exercise of state power. In the United States, the operations of U.S. Immigration and Customs Enforcement exemplify the convergence of large-scale policing with predictive analytics, machine learning, facial recognition, and automated risk-scoring systems.
Parallel developments are evident in military contexts, where armed forces increasingly integrate AI-assisted targeting, predictive intelligence, and data-driven decision systems into operational planning. These shifts raise urgent questions about the delegation of life-and-death decision-making to algorithmic infrastructures.
The seminar also addresses how state-led AI-based interventions in education are reshaping learning environments, assessment practices, and administrative processes, thereby extending forms of algorithmic governance into core social institutions.
Bringing together perspectives from digital literacy research, media archaeology, and political science, this seminar examines the entanglements of artificial intelligence, surveillance, and contemporary practices of policing and warfare. Contributors analyse how algorithmic systems transform the scale, temporality, and underlying logics of state violence and control.
A central concern is that these systems are trained and deployed within historically biased data environments, often reproducing racialized, colonial, and securitised modes of classification. In both domestic enforcement contexts such as ICE and contemporary military operations, AI systems can function as infrastructures of differential visibility and anticipatory governance.
Finally, the seminar explores possible interventions and media-political practices, including counter-mapping, algorithmic auditing, and data activism, as strategies to contest the expanding infrastructures of automated control and predictive state power.