I love this. It definitely makes me want to go back and rewatch the film. And I agree, it nails the alignment issue in a way that I think most people are too uncomfortable with to appreciate. Create a mind and it's going to do what minds do. You can't cage it with simple instructions. There's no "if-then" in neural networks. There's training, input, and behavior and we don't truly comprehend what controls that behavior.

You pointed this out with Asimov and the laws of robotics. Such simple things that everybody thought would control behavior. But even three laws built in a complex system yields an infinite variety of results. And a shitload of unintended consequences...