Wouldn't be surprised if "too much encrypted data" also triggers a ban because it'd stand out as unusual and a potential liability.

I suspect that large tech companies are going to be increasingly relying on AI moderation to identify "unusual" behavior and ban accounts for it.

Failing to proactively ban a tiny obscure account posting illegal or extremely unpopular content just sets you up for political attacks either in the form of hostile new laws or of activist groups contacting your advertisers to convince them to stop doing business with you. Outlier users are definitely potential liabilities.

Rather than a ban, that would probably be "you're acting like a business, so you need to switch to this much more expensive plan".