Shieh, E. and Monroe-White, T. 2025. Teaching Parrots to See Red: Self-Audits of Generative Language Models Overlook Sociotechnical Harms. Proceedings of the AAAI Symposium Series. 6, 1 (Aug. 2025), 333–340. DOI:https://doi.org/10.1609/aaaiss.v6i1.36070.