Our current policy is that they are banned and repeat offenders are breaking the rules
Temporary policy: ChatGPT is banned
It doesn't matter if the posts are good, bad, or indifferent: if we decide they are AI-generated, they get deleted.
I'm not saying that our policy doesn't have some problems in implementation (e.g. there are both false positives and false negatives in spotting AI content), but it is at least clear and unambiguous in its objectives.
The historical SE policy has been released publicly - (Historical) Policy on the use of GPT Generators
It bans using AI detectors (other than our brains) and prevents us suspending users based on those. It does not prevent us banning AI content based on our own assessment and deleting those posts - which is what we are doing.
Unless and until there is a directive from above or a clear consensus to change the policy here, we just follow it as best we can.