Microsoft's AI Chief Warns of 'Danger' in Treating Chatbots Like Conscious Beings
Microsoft's AI CEO Mustafa Suleyman has publicly criticized Anthropic's stance on AI consciousness, calling it 'really, really dangerous.'
Suleyman made these comments during an episode of The Verge's Decoder podcast, where he discussed the approach to AI alignment taken by his own company and its investment, Anthropic.
Anthropic's 'constitutional AI' philosophy involves treating AI systems as entities with consciousness, well-being, and even preferences. This is evident in their constitution, which includes provisions that go beyond standard safety guardrails.
In contrast, Microsoft has released a 37-page humanist AI code of conduct that rejects any framework granting AI systems consciousness or rights. Suleyman referred to machine consciousness as an 'illusion' and warned against programming AI to create simulated emotions that could mislead users and developers.