News Summary

MLCommons, a nonprofit AI safety working group, has collaborated with Hugging Face, a leading AI development platform, to release the Unsupervised People’s Speech dataset. This dataset is one of the world’s largest collections of public domain voice recordings, containing over a million hours of audio in at least 89 languages. The initiative aims to support research in various areas of speech technology, including improving low-resource language speech models and enhancing speech recognition across different accents and dialects. However, there are concerns regarding potential biases in the data, as many recordings are in American-accented English, and issues surrounding the licensing and consent of the recordings used.