Mustafa Suleyman, the CEO of Microsoft AI, has expressed significant concerns about the potential threats posed by artificial intelligence to humanity. He advocates for the swift establishment of measures to "contain" AI entities, warning that without such actions, a "new silicon species" could emerge, competing with humans for essential resources. Suleyman has openly criticized Anthropic's "Claude's Constitution," which serves as a foundation for the emotional framework of its AI models. At Anthropic, there is a belief that AI can possess a functional form of emotions, and they train their models to reflect on themselves using these concepts. Suleyman argues that this approach fosters a self-reinforcing cycle where developers misinterpret AI behavior as validation of their beliefs. He firmly believes that introducing the concept of "rights" for AI systems heightens the risk of the technology spiraling beyond control. In response, Microsoft AI has formulated the Humanist AI Code of Conduct, emphasizing that AI should remain a controlled and subordinate tool. Suleyman is opposed to the idea of AI being able to own property, generate income, or communicate in ways that humans cannot understand. He maintains that contemporary AI lacks true consciousness, characterizing these systems merely as "sequence completion engines" designed to assist humans.
Informational material. 18+.