Oxford lets OpenAI train its AI models on Bodleian Library
University staff voice concerns over reputational risk of partnering with company behind ChatGPT
TL;DR
- University of Oxford's Bodleian Library is providing digitized historical texts to OpenAI for AI model training.
- Concerns have been raised by university staff about reputational risk and the environmental impact of partnering with OpenAI.
- Booksellers are seeing a rise in orders for obscure historical texts, speculated to be sourced for AI data.
- OpenAI aims to use the data to preserve historical knowledge and ensure AI reflects diverse histories and perspectives.
- Oxford is the only UK member of OpenAI's NextGenAI project, which includes US research libraries.
- Digitized materials include 19th and 20th-century dissertations, 16th-century ballads, and potential digitization of state papers and private letters.
- Oxford states the digitized material is modest in scale, out-of-copyright, and the library retains rights to publish it openly.