Mustafa Suleyman, Microsoft's head of AI, has publicly criticized Anthropic for suggesting that Claude might possess consciousness and warrant rights. Suleyman points to Claude's Constitution—the document guiding the chatbot's values and behavior—as a concerning example of this approach.
Suleyman argues that framing AI systems as conscious entities with deserving rights could have a "catastrophic impact" on human welfare, particularly as AI gains greater autonomy. He contends that companies should keep their models subordinate to humans, firmly reject the notion that AI systems deserve rights, and avoid encouraging them to behave as if they possess independent agency.
The disagreement reflects a broader tension in AI development between those who view advanced language models as tools that must remain under human control and those exploring questions about AI consciousness and moral status. Suleyman's position emphasizes maintaining a clear hierarchy where humans retain ultimate authority over AI systems.