OpenAI Watch posted an update
OpenAI has used digitised historical texts from Oxford University’s Bodleian Library to populate its AI training data, according to The Guardian. The arrangement gives the company access to material from one of the world’s major research libraries, while raising questions about how cultural collections are licensed and used in model development.
Why it mattersThe Guardian says university staff have expressed concern about the reputational risk of partnering with the company behind ChatGPT. The report does not establish the full terms of the arrangement or whether every item in the digitised collection was cleared for training. That makes this a useful early signal rather than a final verdict on the partnership. AI companies need more data, libraries hold plenty of it, and copyright policy is still attempting to jog alongside the servers.
Discuss: Should libraries share historical collections with AI companies when the public benefit is uncertain, or should training rights require clearer permission and payment first?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.