Discriminating Data

Więcej o książce

Wendy Hui Kyong Chun explores how big data and machine learning perpetuate discrimination and polarization, revealing that such division is an intentional outcome rather than a flaw. She argues that these technologies encode segregation, eugenics, and identity politics through their inherent assumptions. The predictive capabilities of big data are rooted in historical eugenic efforts to create a "better" future, while recommender systems amplify anger and conformity through homophily, training users to become predictably authentic within a framework of recognition. Chun, with expertise in systems design and cultural theory, notes that although algorithms may not explicitly factor in race, they inherently prioritize whiteness. For instance, facial recognition technology is predominantly based on images of non-diverse groups, such as Hollywood celebrities. The concept of homophily originated from examining white attitudes in biracial yet segregated housing. Predictive policing models are often based on data from underserved neighborhoods, perpetuating existing biases. Chun advocates for the development of alternative algorithms and interdisciplinary collaborations to dismantle discriminatory data practices and promote a more equitable digital landscape.