Microsoft AI CEO says AI threats are real, and Anthropic is making it worse

Microsoft AI CEO Mustafa Suleyman warned that real AI threats are already shaping the safety and regulation debate, and he criticized how Anthropic approaches AI “consciousness.” In an interview context tied to Microsoft’s publication of a 37-page “Humanist AI Code of Conduct,” Suleyman argued alignment is only part of the answer, referencing earlier work that containment is not possible and that proliferation is inevitable. He positioned the conversation around building safer systems rather than relying solely on models doing the “right thing” intrinsically. The discussion also examines whether the AI industry should slow down and how concepts like alignment and containment fit the broader risk picture.