> but purely in the AI.

Which is actually worse and, indeed, makes the argument that we definitely need an A.I. Bill of Rights to protect humans.

The base assumption here, that I'm reading from you, is that morality and intelligence go hand in hand.

Human morality (the "evils" you refer to) are based on human empathy, human philosophically "inherent value" and the human experience.

An intelligence without any of those, nor even the basic nerve-inputs of the physical reality of inhabiting a body, is Blue and Orange Morality at best and complete, perhaps nihilistic, metaphysical solipsism at worse.

Both are a horror.