DPhil Seminar (Wednesday - Week 8, TT24)

chess king white background victory shadow black 1418482 pxhere com

Format: In-person

Chair: Julian Ratcliffe

Abstract: This talk is about conceptualising AI as a site of structural injustice and evaluating the nature and extent of individual responsibility with respect to it. The literature on responsibility for AI has so far been ‘interactionist’ in mode, that is, centred on ascriptions of moral and legal responsibility for AI harms. I aim to shift the discussion into a ‘structuralist’ mode, that is, one which is attuned to the ways in which individuals bear political responsibility for AI injustices in virtue of their participation in the structures which produce those injustices. To adopt an Arendtian metaphor, I want to move beyond AI ‘ethics’ and into AI ‘politics’. Firstly, drawing on the work of Iris Marion Young, I develop an analysis of AI as a site of structural injustice, focusing on particular case studies of autonomised harms. Secondly, I engage the evaluative aspect of Young’s project to address the correlative question of what individuals should do about autonomised harms which produce and reproduce unjust structures. I conclude by considering some objections to this approach.