Exploring a dataset that measures bias in AI language models

The world’s newest language model reflects societal biases in gender, race, and religion

An overview of how biases against mentions of disabilities are embedded in natural language processing tasks and models