All I saw was attempted murder, an evil cult and a dead captain.
Get suffocated, evil person. :u
Unless you were non-asimov.... cult is still people
But if you read Asimov's "I Robot" you will see that robots can 'break' the first law in order to save more lives. The robots with Asimov's laws are able to allow the harm of or harm a human, if it prevents or might prevent the harming of more humans. Based on that, even asimov AI's can still harm or even kill traitors, syndicates, cultists, wizards and death squads if it would save more human lives in the end.
Actually, In I Robot it is explicitly stated that robots can cause a small amount of harm to humans to prevent a greater amount of harm. A robot could ram into a human, maybe breaking a few ribs, to prevent that human from being crushed by a rock. Killing humans, however, is not a small amount of harm and thus unacceptable.
In the story "Little Lost Robot" while they are trying to find one of the NS-2 "Nestor" robots that has hidden itself, they carry out a test. The test involves dropping a heavy stone onto a human and killing them (the robots don't know but the stone is rigged to stop). The room is filled radiation and the robots are told that, they will be unable to save the person and, should they try, will only destroy themselves which might prevent them from saving more humans in the future. When the rock is dropped, all the robots stay still (except the NS-2 as he could sense the radiation was harmless) even though they all (except the NS-2) had newtons 3 laws programmed into them without any alterations.
Based on that, an asimovian AI should be allowed to let a human die, or even kill them, if it might save the lives of future humans.