This is a great answer, and I think it would be relatively straightforward to create filters for potential bot/spammer accounts that used this strategy (ex: high comment frequency, using similar text patterns and engaging mostly in an unidirectional way) and make it available for users to enable it.