Séminaire de Philosphie et IA
Nous aurons le plaisir d'écouter Carina Prunkl (philosophy, INRIA)
Titre : AI as Bureaucratic Decision-Making
Résumé : AI-driven decision-making is frequently evaluated against individual human judgment, with debates centering on its efficiency, consistency, and potential biases. In this talk, I challenge this framing, arguing that the more appropriate comparison class for AI-driven decision-making is not individual human judgment, but bureaucratic decision-making. Both AI-driven and bureaucratic decisions rely on quantification, rule-based standardisation, depersonalised authority, and task specialisation. Both have been praised for their consistency and impersonality, yet also criticised for their lack of context-sensitivity, lack of transparency, and the removal (or redistribution) of human discretion. This shift in perspective provides a new way to examine questions of accountability and legitimacy in algorithmic decision-making, drawing on the rich historical, theoretical, and practical scholarship on bureaucracies
ID de réunion : 925 0375 0086
Code secret : 098704