The main part of the policy seems to be the following:

"We've reminded moderators that suspensions (and typically mod messages as well) are for real, verifiable malfeasance only, and should not be enacted on the basis of hunches, guesses, intuition, or unverified heuristics."

While I understand that StackOverflow does not want false positives, the only kind of AI generated text that is truly verifiable, is the one that still contains "As an AI language model..."