A new paper by MScAC alumna Aparna Balagopalan demonstrates why labelling data with normative prompts can yield better outcomes in machine learning models. Its co-authors include Assistant Professor, Status-Only Marzyeh Ghassemi and CS graduate student David Madras.