How many bots are there?
How many bots are "abusive"?
How many bots are " useful"?
Until this is known, we won't even know what we're dealing with.
Can they be identified and classified?
I believe a good start is to provide a way for humans to identify"bottish" actions of an account, then let humans decide what to do about them.
I don't advocate giving bots right (and I never have), but I do believe humans have the right to run bots.
When and if it is decided that the operations of an account are abusive, then the account can be hammered to oblivion.
RE: The Difference Between Promotion and Curation, and Killing All Bots