> Suleyman argues AI models may become even harder to control if they're told to think they have feelings and rights: "Frankly, with this additional baggage, I think it would make them a catastrophic threat to human civilisation," Suleyman says.

This feels like a villain's reasoning, and I'm kind of astounded he's just saying it out loud.