For me personally... I think it's when they are a bot made for nefarious purposes to the detriment of the network/protocol and it's entire community. I didn't mind the follow bots for over a year now....
That’s not my question. In order to create a bot filter that works reliably, I need to know how to define them with logic. What specific criteria defines a bot that I can use to detect them just by scanning their profile?
1