Harvard University, in conjunction with Google, has released a dataset of a million public domain books to train the next generation of AI. The books span genres, languages, and authors such as ...