Can the AI harm harmful non humans?

[CPTANT]

Bottom post of the previous page:

What is acceptable escalation of the AI versus harmful non humans?

What is the reasonable response for non humans doing the following to humans: light harm, significant harm, attempted murder, murder and mass murder?

Is there a difference between asimov and other lawsets defining groups of humans and non humans? I get the feeling lots of people see asimov as the "good" guy even when dealing with non humans, but I completely disagree with this interpretation. Asimov is a neutral force and only distinquishes two parties: humans and non humans. The only fundamental difference with a one human/syndicate lawset is who is defined human.

[Grazyn]

Yeah if a non-human officer is stupid enough to harm a human in front of a borg, he deserves the logical outcome.

[Hulkamania]

An AI/Silicon should be acting within the interests of the crew and preventing harm whenever possible. If someone is an active source of harm, that should be acted against.

However, rule 1 still applies in these circumstances, much like it would for a purged AI. You have a loosened escalation when it comes to doing whatever it takes to protect humans, but that should not include outright murder/round removal if there is a non-violent or less dickish way to do so.

Obviously this only applies to standard asimov, when law changes come into play its best to just ahelp if you're ever unsure about a course of action you're considering. (Plus if you have to ahelp a course of action in the first place you're probably incorrect for doing it, although this is more of a general rule of thumb)