Explore a 21-minute conference talk from the FAccT 2021 virtual event that delves into the challenges of fair classification in the presence of group-dependent label noise. Presented by J. Wang, Y. Liu, and C. Levy as part of the Research Track, this talk examines the impact of biased data labeling on machine learning models and proposes solutions to mitigate unfairness in classification tasks. Learn about the researchers' approach to addressing this critical issue in AI ethics and fairness, and gain insights into potential strategies for improving the accuracy and equity of machine learning systems when dealing with noisy, group-dependent labels.