Microsoft AI leader warns of risks from autonomous artificial intelligence systems

Mustafa Suleyman, who leads AI development at Microsoft, has warned that AI could become a kind of competing species if companies do not control it carefully. He made these comments in a BBC interview this week.

Suleyman said that today's AI systems do not think or feel. They simply predict and complete sequences of text. He criticised the company Anthropic because he believes it treats its AI, Claude, as if it were conscious. He said this gives people a false idea of what AI really is, and that this could be very harmful.

Microsoft has published a set of guidelines called a Humanist AI Code of Conduct. It describes how AI should remain under human control. The company also created a superintelligence team in October 2025.