I asked a manager about this, the policy is that we do not need to differentiate between bots and people who sound similar to bots: both are considered low quality content/engagement. Delete them.
Seems like wherever they delete bots, they will in the end, delete human beings.