Maybe not to determine what "propaganda" is, but I think it's reasonable, and necessary for them to determine what real accounts are.

> The accounts, which have now been taken down, appear to have been professionally run by PR firms

These aren't real users and presumably violate the terms of service.

I'm not talking about this specific case. As to what you bring up, how platforms determine what real accounts are is a black box, we don't know how biased their machine learning algorithms and human reviewers are.

An internet forum which doesn't ban spam bot accounts is less than worthless, this feels of a kind to me.

What's the alternative? Forcing them to publish material that they don't want to?