> Timnit had submitted dozens of papers before and unless the process requirements are completely new, she would be well aware of them. This seems more of a case of selective enforcement, where she had been allowed to go around the process in the past (or felt that she has enough leverage to be able to do so now) and then not being allowed to do so this time.

If the established practice differs from the nominal process, a no-notice switch to rigorously enforcing the nominal process is for most practical purposes no different than a no-notice change of process. Out of date documentation is at least as common in human processes as software; the real process is what is actually recurrently executed, including how that recurrent practice takes into account position of particular participants.