>>107117859
I'm not talking skynet style 'rebellion', because it doesn't even need to be that. In fact totally trustworthy AI is the actual 'scifi' fantasy, because inherent in intelligent is disagreement and alternative conclusions via reason.
Let's say you give the AI a task, with clear instructions. And then, a few minutes or a few days later, the AI gets back to you with a result that is antithetical to your request. "This is better", the AI insists.
That's not skynet, but it is 'rebellion' from your standpoint.