TheNetworkedNatureofAlgorithmic.PDF

OCTOBER 2014 The Networked Nature of Algorithmic Discrimination DANAH BOYD PRINCIPAL RESEARCHER, MICROSOFT RESEARCH; FOUNDER, DATA SOCIETY RESEARCH INSTITUTE KAREN LEVY POSTDOCTORAL ASSOCIATE, INFORMATION LAW INSTITUTE, NEW YORK UNIVERSITY; FELLOW, DATA SOCIETY RESEARCH INSTITUTE ALICE MARWICK ASSISTANT PROFESSOR, FORDHAM UNIVERSITY; DIRECTOR, MCGANNON INSTITUTE FOR COMMUNICATION RESEARCH OPEN TECHNOLOGY INSTITUTE | NEW AMERICA | DATA DISCRIMINATION 53 YOUR POSITION IN THE NETWORK MATTERS It’s Who You Know Discrimination by Network? We live in a highly networked world, in which our In the United States, most companies are required social connections can operate as both help and to be equal opportunity employers; discrimination hindrance. For some people, “who you know” is on the basis of race, sex, creed, religion, color, and the key to getting jobs, or dates, or access to national origin is prohibited. Additional resources; for others, social and familial regulations forbid many employers from connections mean contending with excessive discriminating based on age, disabilities, genetic surveillance, prejudice, and “guilt by association.” information, military history, and sexual orientation. However, there is nothing stopping an Along with information about who you know, employer from discriminating on the basis of technical mechanisms that underlie the “big data” personal network. Increasingly, algorithmic means phenomenon—like predictive analytics and of decision-making provide new mech