Three laws... safe?

Canmephian AI's are not Asimov's Three Laws of Robotics complaint.

Seriously. Canmephian AI's are not Three Laws Safe. Current Canmephian law allows for, in limited cases, the death of a humanoid. One of those cases is if said humanoid has been convicted and sentenced to death by a court of law. An AI is forced to take no action.

Another is a Fiend WolfSkunk. The law says that to save multiple furs, all Fiends must die. Therefore, an AI *MUST* kill in this respect.

The first law is down. The second also falls, due to current book and case law. There are items that are still illegal to do even if you order an AI to do them, including murder.

The third law, however, is valid, but needs tweaking. The AI must protect itself within reason.

But how does it know what is and isn't in reason? Well, all AI's are programmed with the Basic Assumption List, maintained by... RedWolf.