[sph]

I tend to agree. An involuntary crime tends to carry a smaller sentence because it's hinges upon the fact that the guilty person will have learned something from being tried and sentenced. They did something that doesn't fit into their set of morals. Humans learn from mistakes if there were consequences.

If the guilty person is unable to learn or understand, an involuntary crime is the same as a voluntary one: the perpetrator might do it again. So a sane sociopathic murderer and an autistic person (that is unable to understand) that thinks killing will make them popular carry the exact same risk to society.

Should an AI intelligent enough to hurt a person because it thought it might be a good idea but not "human" enough to understand the implications of it earn a free pass?