AI companies that are eager to access the library’s digital archives — and the 185 petabytes of data stored within it — to develop and train their most advanced AI models. “We know that we ...