Systemic Algorithms and Society looks at issues of computational bias in the contexts of cultural works, metaphors of magic and mathematics in tech culture, and workplace psychometrics.
The output of computational models is directly tied not only to their inputs but to the relationships and assumptions embedded in their model design, many of which are of a social and cultural, rather than physical and mathematical, nature. How do human biases make their way into these data models, and what new strategies have been proposed to overcome bias in computed products?
Scholars and students from many backgrounds, as well as policy makers, journalists, and the general reading public will find a multidisciplinary approach to inquiry into algorithmic bias encompassing research from Communication, Art, and New Media.
This book is one of the titles in Routledge's "Algorithms and Society" series. Other titles in the series include Digital Totalitarianism, Privacy, and Deep Fakes. This fairly small volume contains three chapters of length ~20 pages.
- "From 'Diversity' to 'Discoverability': Platform Economy, Algorithms and the Transformations of Cultural Policies" (by Christophe Magis). [Analyzes how the movement away from earlier critical studies of the global cultural economy has produced the weak concept of 'diversity.']
- "Modern Mathemagics: Values and Biases in Tech Culture" (by Jakob Svensson). [An attempt to understand tech culture, its values, and its biases through the metaphor of magic.]
- "Reading the Cards: Critical Chatbots, Tarot and Drawing as an Epistemological Repositioning to Defend Against the Neoliberal Structures of Art Education" (by Eleanor Dare and Dylan Yamada-Rice). [A critique of the neoliberal structures in universities, which undervalues alternative ways of knowing, such as making, drawing, and experimenting with materials.]
I found the coverage of topics somewhat haphazard and the depth of treatment rather disappointing.