I find it entirely implausible that facebook, with their vast troves of data and an army of the very best computer scientists money can buy, is incapable of reliably identifying posts and users responsible for mass manipulation on their platform. Unwilling? Sure. Unable? Not a chance.
With the current logic that Facebook is using, "manipulation" can mean any behavior that they don't endorse. For example someone is "manipulating" their network by posting anti Biden memes.
I recognize that I am extrapolating real group behavior to real individual behavior, but I think that extrapolation is warranted given that we've just made the jump from group bot behaviour to group people behavior.