Skip to content
Tech · San Francisco

Oxford's Bodleian texts were used to train OpenAI models, documents show

Material digitised under a 2025 partnership went into OpenAI's training set, according to internal documents reported by The Guardian.

·
The Divinity School in Oxford's Bodleian Library
The Divinity School in Oxford's Bodleian Library.

The University of Oxford allowed OpenAI to train its AI models on historical texts from the Bodleian Library, The Guardian reported on Saturday, citing internal documents that say the digitised material was used to populate the OpenAI training set.

Oxford announced a partnership with the San Francisco company in March 2025 to digitise texts from the library using OpenAI software, saying the project would make the collection more widely available to students and researchers. That announcement did not say the material would be used for model training.

An OpenAI spokesperson said the company was proud to help ensure today's AI models preserve historical knowledge for the future.

The report highlights how AI developers are turning to universities and archives for new training data, and why publishers, libraries and authors are paying closer attention to the terms of digitisation deals.

From ARQS Sponsored

Readers interested in the creative economy can explore AI receptionist, ARQS Store, AI Academy courses and Northline Technology Institute. In the ARQS Store, ARQS Website Concierge Setup, Prompt Architecture for Business, Online Academy-in-a-Box are among the newest tools available.