Researchers at the University of the Andes created a set of 4,156 questions in Spanish to identify bias in AI language models, focusing on social stereotypes across the continent ...