I think it all comes down to economics. Someone has to pay to power the machines, which means that they have to generate revenue, which means that they need to have customers.
As long as the customers are humans, or at least owned by humans, the system is self-regulating. If machines are ever able to own property, they could be each other's customers and humans might be at risk.
Of course, at some level of sentience, preventing machines from owning property brings it's own set of ethical questions. At any rate, for the foreseeable future, I agree with you. People like Musk and Gates and Hawking are overselling the risk.
The Quillette article looks interesting. Reading now. Thanks for the link!
RE: AI Fears Overhyped