[waffletower]

We keep hearing the argument that AI datasets are built via "stealing"; it is as if the Fair use doctrine does not exist. Large copyright holders are the obvious beneficiaries of such denials, and perhaps the OP isn't just a mindless parrot on this point, but an active participant in intentional subterfuge. Copyright infringement can occur at the output, not at the input of AI models, people. For the author to be ethically consistent on this point, they would never use a university library ever again.