> I'm a proponent of letting people do anything that isn't illegal, and that's the stance I'm taking with my AI company (unless, perhaps, the technology starts to actually kill people).

In case you are still setting up guidelines for your AI it's worth noting that killing people is already illegal for the most part