The Center for Critical Race + Digital Studies (CR+DS) is at the forefront of understanding inherent bias in machine learning and algorithms that shape our digital worlds. CR+DS, an affiliate center of the Institute of Human Development and Social Change (IHDSC), is a network of prominent, public scholars of color who produce research, distribute knowledge, and convene stakeholders at the intersections of race and technology. Their affiliates include social scientists, artists, computer scientists, cultural critics, lawyers, policymakers, ethicists, journalists who understand how digital media technology manifests racialized power relationships throughout our societies.
They have recently released two incredible resources that dispel myths that machine learning and algorithms are objective and bias-free and refute false claims of race- and value-neutral operationalizations in the digital space. Learn more and download the complete package of resources at the CR+DS website.
A People’s Guide to Finding Algorithmic Discrimination.
This practical guide was led by CR+DS affiliates Meredith Broussard and written by Avriel Epps-Darling, in partnership with Rumman Chowdhury and the team at ParityAI. The guide features four critical lessons: (1) What is a machine learning algorithm?; (2) Algorithmic Bias?; (3) What types of Bias?; and (4) How do these emerge in real-world contexts? It features several demonstrations to help you get hands-on practice using state-of-the-art detection and debiasing tools, along with the underlying data and code, with examples of bias in the domains of employment/hiring, credit worthiness, risk assessments, predicting diabetes, facial recognition, and predicting student performance. This is a one-of-its-kind guide that is a resource for those with no familiarity with coding— and those who do— and provides extensive links to the major pioneers that have brought us this work.
Targeted Ads: The Infrastructure for Algorithmic Discrimination
The Targeted Ads research report led by CR+DS affiliate Matthew Bui, University of Southern California PhD student Ho-Chun Herbert Chang, with advisement from Charlton McIlwain, features novel data, methods and a toolkit for auditing algorithmic discrimination in targeted ads. Among others, the report concludes that “Online targeted ads uphold, produce, and recreate racially discriminatory infrastructures within everyday life.” Download the full report.
Learn more about CR+DS and the incredible work they are doing to bring awareness to algorithmic bias and their fight for more inclusive tech workplaces.
