Comment by blensor

Who says the counter agents don't decide to shut down a powerplant to end an attack it's otherwise unable to contain.

If counter AIs have strict safeguards they are disadvantaged by design, if they don't have them they are potentially equally dangerous as the attacker