Machine learning purports to make accurate predictions and decisions about everything from prison recidivism rates to cancer diagnoses to mortgage approvals. But time and time again, scholars, activists, and journalists have demonstrated that machine learning algorithms often digitize and replicate inaccuracies, historical prejudices, and institutional harms. With so much on the line, machine learning models and the data used to train them may deserve more scrutiny.

Join us as we explore the challenges of machine learning ethics with Abeba Birhane, Ph.D. candidate in cognitive science at University College Dublin, and Noopur Raval, postdoctoral research fellow at the AI Now Institute at New York University.