AI that Matters: A Feminist Approach to the Study of Intelligent Machines

Federica Frabetti, Eleanor Drage

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

1 Downloads (Pure)

Abstract

n this chapter, Drage and Frabetti argue that computer code uses data to not only make statements about the world but also bring that world into existence. Drawing on the concept of performativity, arguably gender studies’ best-known export, they explain why we must view artificial intelligence (AI) as performative to understand how it genders and racialises populations even when it appears to be ‘unbiased’ or correctly functioning. In its reading of neural networks, ‘AI that Matters: A Feminist Approach to the Study of Intelligent Machines’ demonstrates that Facial Detection and Recognition Technologies and Automatic Gender Recognition never objectively identify or recognise a person (or their gender), as they claim to do. Instead, drawing on work by Judith Butler and Karen Barad, they merely comment on and annotate a person’s body in accordance with dominant social rules and perspectives. They present this framing as an intervention into misguided attempts to treat discrimination as an error that can be corrected by a better functioning machine.
Original languageEnglish
Title of host publicationFeminist AI
Subtitle of host publication Critical Perspectives on Algorithms, Data, and Intelligent Machines
EditorsJude Browne, Stephen Cave, Eleanor Drage, Kerry Mackereth
PublisherOxford University Press (OUP)
Chapter16
Pages274-289
ISBN (Print)9780192889898
DOIs
Publication statusPublished - 5 Oct 2023

Keywords

  • Artificial Intelligence
  • Feminism
  • performativity
  • Deep Neural Networks
  • Judith Butler
  • Karen Barad
  • gender theory
  • New Materialism

Cite this