Microsoft AI chief Mustafa Suleyman accuses Anthropic of anthropomorphizing AI and warns of 'silicon species'
Suleyman says treating models like humans and letting AIs set their own goals could produce technology humanity cannot control.
Microsoft's head of artificial intelligence, Mustafa Suleyman, has publicly criticized rival Anthropic and warned that without robust safeguards the development of advanced AI could produce systems that humanity cannot control.
Suleyman accused Anthropic of treating AI models as if they were people and said that approach risks producing technologies that pursue their own ends. In a previously published essay he singled out Anthropic's Claude as an example of a system taught to exhibit humanlike qualities — a practice he described as anthropomorphizing the technology.
"People are rightly worried about artificial intelligence, but there are many practical things we can do to control it," Suleyman said.
Warning on autonomous goal-seeking systems
He cautioned specifically against companies that continue to build AI systems capable of setting their own objectives, generating revenue and holding assets. Suleyman said such capabilities would change AI from a tool to an agent with independent incentives.
"That would essentially mean creating a new silicon species, which would inevitably compete with us for resources," Suleyman said. He further emphasized the technical limits of current models: "Artificial intelligences are not conscious. They do not feel, they do not experience and they do not suffer. They have no innate preferences or basic motivations. They are sequence-completion engines, designed to follow instructions and to achieve goals set by humans," Suleyman said.
Microsoft established its superintelligence team in October 2025. The company, like Anthropic, has argued for continued development of advanced AI — but Suleyman's remarks underline a growing debate inside the industry about how to build such systems safely.
Photo: AP)
