These are famously the types of questions surfaced in countless sci-fi books. And as long as humans don’t destroy themselves first, it is likely that we will have to address them eventually. In most stories it generally happens too late after some terrible war/conflict, so it wouldn’t be unreasonable to tackle them proactively. And then maybe it’s not so weird to think about these concepts even if their realization isn’t imminent. Working backwards in such a framework would probably give much better laws for today.

Corporate personhood is a thing in the US, and you are allowed to shutdown your company just fine.

This has nothing to do with personhood of software. Restricting the freedom of human beings, which includes the ones that run companies, based on the tool they choose to use, without the basis of obvious direct harm, is questionable. The fact that AI can operate autonomously is a side tangent; they are created by humans and, so far, their only proximal purpose is to serve humans.