Can AI Be Trusted?

[LegitHyperbole]

Fantastic thread. In it's current form, yes and no. It can hallucinate and come up with all sort of nonsense but will it secretly kill us, no. Post singularity Ai, I would hope yes but how can we know? It depends on how it's developed, what the nature of phenomalogical consciousness is among many other very deep and nuanced questions we don't have the answer for. In all likelihood we would be building a psychopathic intelligence because it would not have the millions of years of evolution baked into its DNA. Would you trust an AGI that doesn't find kittens and puppies cute or find the sunset beautiful? Would you trust a super intelligence that can't feel love or friendship.