The deep border

Amoore, Louise

Abstract

Deep neural network algorithms are becoming intimately involved in the politics of the border, and are themselves bordering devices in that they classify, divide and demarcate boundaries in data. Deep learning involves much more than the deployment of technologies at the border, and is reordering what the border means, how the boundaries of political community can be imagined. Where the biometric border rendered the border mobile through its inscription in the body, the deep border generates the racialized body in novel forms that extend the reach of state violence. The deep border is written through the machine learning models that make the world in their own image – as clusters of attributes and feature spaces from which data examples can be drawn. The ‘depth’ that becomes imaginable in computer science models of the indefinite multiplication of layers in a neural network begins to resonate with state desires for a reach into the attributes of population. The border is spatially reimagined as a set of always possible functions, features, and clusters – as a ‘line of best fit’ where the fraught politics of the border can be condensed and resolved.

Citation

Amoore, L. (2024). The deep border. Political Geography, 109, Article 102547. https://doi.org/10.1016/j.polgeo.2021.102547

• Journal Article Type: Article
• Acceptance Date: Nov 9, 2021
• Online Publication Date: Nov 25, 2021
• Publication Date: 2024-03
• Deposit Date: Jan 26, 2022
• Publicly Available Date: Jan 28, 2022
• Journal: Political Geography
• Print ISSN: 0962-6298
• Electronic ISSN: 1873-5096
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 109
• Article Number: 102547
• DOI: https://doi.org/10.1016/j.polgeo.2021.102547
• Public URL: https://durham-repository.worktribe.com/output/1215811