Antidiscrimination Law and Algorithmic Bias

Human decision making is increasingly being displaced by algorithms. Judges sentence defendants based on "risk scores;" regulators take enforcement actions based on predicted violations; advertisers target materials based on demographic attributes; and employers evaluate applicants and employees based on machine-learned models. A predominant concern with the rise of such algorithmic decision making (machine learning or artificial intelligence) is that it may replicate or exacerbate human bias. Algorithms might discriminate, for instance, based on race or gender. This course surveys the legal principles for assessing bias of algorithms, examines emerging techniques for how to design and assess bias of algorithms, and assesses how antidiscrimination law and the design of algorithms may need to evolve to account for the potential emergence of machine bias. Admission is by consent of instructor and is limited to 20 students. Student assessment is based on class participation, response papers, and a final project. CONSENT APPLICATION: To apply for this course, students must complete and submit a Consent Application Form available on the SLS website (https://law.stanford.edu/education/courses/consent-of-instructor-forms/). See Consent Application Form for instructions and submission deadline.