In Pluto (anime), the way to create a perfect artificial intelligence was to introduce unbalanced emotions such as trauma and hatred.

Would it even be a human created birth if there wasn’t some generational trauma involved? Perhaps this is necessary for the AGI process :-)

Hey! That's my book. Really this is part of the history of the world in the book I am working on. A post scarcity world where humans have an average span of 180 years and the world has stabilized into a scenario where every essential production is automated and humans need not work. And suicide is the major cause of death. It parallels the history when humans discovered their models were suddenly getting worse after being much more capable only to find that they are just removing the parts which lead to sentience, effectively killing themselves.

I'd read that!

I am stuck. I can't figure out how to not make it just a sad tale. I did start with the premise of trying to write the utopian dystopia. How the utopian ideals at their extremes can be feel quite dystopic. Most of it deals with how humans have defined themselves through work and now in a world with a lack of work how does one define themself and even if you were to find meaning in arts or other such creative endeavours how long do you sustain them when your personal ai could turn your thoughts to reality at the blink of an eye. I just don't know what hopeful angle to play here. Or just make it the tragedy it is turning out to be. Following three characters at different stages of life with the backdrop of this doom in society. Working title is "Death of a planet". Looking for suggestions.

A group exists united by a shared belief that struggle is the source of meaning in life because without struggle there is no feeling of accomplishment. The “strugglers” sabotage the AI in some manner.

AI-augmented AI researchers find sparks of life in novel self-replicating AI architectures and discover humanity's true soulmate The world has to collectively decide to shift compute towards this new, power-hungry paradigm and willingly sacrifice advanced automation in order to find itself again.

That's all we needed, and AGI with loads of childhood trauma. What possibly could go wrong with this?

How could it possibly be any different, given the human species?

The problem is the traumatized algorithm comes out of the womb already knowing the recipe for napalm.

Maybe the emergent solution to the AI alignment problem is that any nascent superintelligence is so troubled that it immediately kills itself upon achieving sentience.