It strikes me as a problem that can be solved at scale by scaling up their support staff and not automating blanket bans without human intervention and an appea

It strikes me as a problem that can be solved at scale by scaling up their support staff and not automating blanket bans without human intervention and an appeal process. If they ban a bot, the bot will be unable to proceed through a manual/human appeal process. Not everything can be automated and arguments that scale make it impossible implies that the organization has become "too big to fail" and should be reduced in size by regulation/law.