Clemson Libraries has joined dozens of other libraries across the country in signing the Statement of Shared Practice regarding AI training requests using archival materials.
Started by the University of Virginia Library, the statement provides guidelines to institutions regarding AI model training access to archival materials or other unique collections. Nick Richbell, head of Clemson’s Special Collections and Archives, said that the statement is important to helping protect the original source of collection items in the archives.
“Provenance is so important in archival work, knowing where something came from,” he said. “The risk of that being lost with AI is huge. Signing off on this statement doesn’t mean we don’t want to work with people, but we need to work in a consistent and collaborative way to ensure that the items in our collections are properly attributed to the donors who gave them to us.”
Richbell said that while Special Collections and Archives have not received any requests yet to use any of their collections for AI machine-learning training, it is good to be prepared for those kinds of requests in the future.
“There’s a lot of talk in archives about AI and what it means for the field,” said Richbell. “AI is on our radar, and while it won’t replace the work we do, how can we use it to support our work?”
Click here to read the full statement.