De-biasing language

Looking for more information on bias and other pitfalls in AI? Check out the "Ethics, Privacy, and Security" sessions at the AI Conference in New York, April 15–18, 2019. In a recent paper, Hila Gonen and Yoav Goldberg argue that methods for de-biasing language models aren't effective; they make bias less apparent, but don't actually remove it. De-biasing might even make bias more dangerous by hiding it, rather than leaving it out in the open. The toughest problems are often the ones you only think you've solved. Language models are based on "word embeddings," which are essentially lists of word combinations derived from human language.