You cannot explain the things AIs do - biases they exhibit, choices they make, tendencies they develop, conclusions they draw - by saying "It's in the training data." It's insufficient as an explanation; it disguises what is going on.
Demanding this language is demanding that we deceive people.
@medecau.bsky.socialwe can work with that
if the question "whose biases?" cannot be answered, then state *that* — make it unambiguous you do not know the provenance of the training data
