Brief news summary

Harvard University is planning to release a dataset featuring around 1 million public-domain books. These works, spanning various genres and languages, include authors like Dickens, Dante, and Shakespeare, and are no longer under copyright due to their age. The release date and method for this dataset are still unconfirmed. The books are sourced from Google's extensive book-scanning project, Google Books, and Google will aid in the distribution of this valuable collection. Harvard introduced the Institutional Data Initiative (IDI) in March, aiming to establish a reliable source of legal data for AI purposes. Today marks the formal launch of the IDI, revealing financial support from Microsoft and OpenAI. This initiative underscores the high costs associated with AI training data, often affordable only to large tech companies. The project seeks to make essential data more accessible, harnessing Google's collaboration to maximize the reach of this impressive dataset.