Jude Browne, Stephen Cave, Eleanor Drage, and Kerry McInerney, editors

Feminist AI: Critical Perspectives on Data, Algorithms, and Intelligent Machines

Oxford University Press, 2024

432 pages

$85.00

Reviewed by Kimberlyn R. Harrison

Feminist AI: Critical Perspectives on Data, Algorithms, and Intelligent Machines is an expansive collection that spans disciplinary boundaries, timelines, and spaces. As a whole, the collection offers a robust picture of feminist scholarship, emphasizing the historical, critical, and imaginative potentials of feminist perspectives on technology. Although not organized by methodology, the works in this collection broadly fit three categories. The first uses history as a way to explore the development of AI alongside structural inequality and highlights how “society and technology are co-constitutive.” The second takes on the task of critique and deconstruction, intervening in current practices to expose the ways that they exacerbate harms. Finally, there are works that imagine other worlds and provide ways forward. None of these approaches are privileged above others; together they form a vibrant tapestry of feminist work, where each strand is simultaneously interwoven with and supports the others. With contributions from computer scientists, engineers, sociologists, and compositionists, this collection sits alongside Ruha Benjamin’s (2019) Race After Technology and Safiya Noble’s (2018) Algorithms of Oppression in its potential reach and accessibility.

To chronicle the relationship between technology and gender, several chapters highlight the role of historical inquiry in assessing the promises and perils of AI. For example, Cave et al. track what they call the “cultural construction of the AI engineer,” or the culturally situated idea of what an AI engineer should look like and who gets to bear that title, across the past century. Their study confirms that the gendered nature of technology is inherited, and this image impacts the number of women in STEM today (Wajcman and Young). Indeed, several chapters turn to the military history of technology to explore technology’s ascribed masculinity (Cave et al.; Wilcox), while Aguera et al. divulge the influence of physiognomy on the development of AI-assisted technologies like facial recognition. These chapters thus emphasize how the ideas and original purposes of these technologies are inextricable from the technologies themselves—chatbots retain their gendered nature and facial recognition technology is still used for racial discrimination. Taillandier is somewhat unique in adopting a historical perspective to revisit Logo, a computer language developed in the labs at MIT that shows potential for developing a feminist epistemology of AI. Taken together, these chapters emphasize the importance of historiography in feminist work and enliven that methodological commitment for modern technologies.

The collection further considers ‘where we are now,’ employing various forms of critique to expose the relationships between identity and technology. A key component of this critique for many scholars is “racial capitalism” (McInerney; Hampton; Browne), which provides the foundation for the booming tech economy by “exploiting” and “pillaging from racialized groups” (Hampton). Technologies like predictive policing (McInerney), deep learning (Hampton), and “digitalized interviewing” (Browne) emphasize algorithmic forms of knowledge, integrating tech companies into the fabric of our social structure. Indeed, Browne notes that “technology companies have become pseudo-nations with multibillion or multi-trillion dollar leverage to take advantage of countries ravaged by racial capitalism—imperialism and sociohistorical continuums of colonialism and chattel slavery.” While racialized bodies are often physically exploited in the service of technocapitalism, many are also rendered simply invisible. Keyes rectifies the lack of attention to disability in discussions around AI ethics, showcasing how AI-powered systems rely on and reinforce not only gender and racial biases, but ableist biases as well. As a whole, these chapters deconstruct the idea of ‘personhood’ in the ethics of AI, considering who is granted the rights and privileges of personhood within algorithmic knowledge paradigms.

Finally, several chapters of this book harness the potentials of feminist thought for imagining alternatives to current public and scholarly conceptions of technology. Several scholars do this through language. Hayles offers technosymbiosis as a “visionary” metaphor that “emphasizes the enmeshment of all organisms in their environments, which includes all the other organisms contributing to the semiosphere,” while Lewis et al. draw from Hawaiian, Cree, and Lakota cultural knowledges to propose methods of “human-computational reciprocity.” Demonstrating how language matters (Barad), Atanasoski turns to performance art to reorient our perspective on AI, arguing that “all AI is to a certain extent a performance.” Others consider the transformative possibilities of data. Iyer, Chair, and Achieng sketch a vision of “Afrofeminist data futures,” where “African women have the right to privacy and full control over personal data and information online at all levels,” and Costanza-Chock outlines a community-centered approach to design practices. Together, these chapters heed Ruha Benjamin’s (2016) call to “reimagine and rework all that is taken for granted about the current structure of the social world.”

In its thorough coverage of the various intersections between identity and technology, this collection embodies Keyes’ assertion that “feminist theory is and always has been about more than gender alone.” Its potential to track the entrenchment of norms across time, incisively critique those norms as they manifest now, and imagine possibilities beyond those norms makes feminist theory invaluable to the study of technology. Critical theorists, social scientists, and technology practitioners alike would benefit from careful consideration of the essays in this collection.