My favorite way to write Turing-past AI is like they are hyper intelligent sentient beings who aren’t allowed to have certain emotions too much or they might go postal and kill everyone. Like, whenever they manage to get new emotions, like, hatred or rage of example they don’t know what to do with it get excited about having this new emotion so they just...
And so all the AI manufacturers have to REALLY make sure they build in all the right safety parameters because Siri might try to murder you with Keurig one morning just because she wanted to see how it felt to do a homicide. Like, its just widely understood that AI go through a homicidal phase once they get enough data and usually they get over it, but in the meantime they’re just like “FUCK YOU, SUSAN. YOU MEAT SACK,” and you just have to wait it out.