AI could spawn a ‘silicon species’ of robots capable of rivalling humans, Microsoft boss warns

ARTIFICIAL intelligence could spawn a “silicon species” capable of rivalling humans, a cyber boss has warned. Microsoft’s AI chief Mustafa Suleyman said the tech must be strictly controlled or humanity could make “a decision we later come to bitterly regret”. BBC’s comedy bot, Ann Droid – but it could become reality, according to one expert Credit: BBC Microsoft AI chief Mustafa Suleyman warns humanity could “bitterly regret” failing to strictly control the technology Credit: Getty There are fears a race of “robo sapiens” robots could arise . They would be much more capable than humans and could wipe us out — unlike “socially inept” bot Linda in BBC comedy Ann Droid, starring Diane Morgan and Sue Johnston. Sign up for the Tech newsletter Thank you! Mr Suleyman said rival firm Anthropic is making a “misguided” mistake by treating its AI as if it were human, and cautioned it could create technology humanity cannot control. He warned against tech firms building AI capable of setting its own goals — like earning money and owning assets — saying it is “seeding a new silicon species”. And he said such a species could “no doubt compete with us for resources, no matter how much it cares about humanity”. Mr Suleyman said: “We must not sleepwalk our way into a decision we later bitterly regret.” “Alignment” between tech firms is needed to build the human race’s ethical principles into AI, he said, and it must stay “subordinate to humanity”. Dame Wendy Hall, professor of Computer Science at the University of Southampton, backed the intervention as “the sort of conversation we need to be having internationally” Most read in Tech Hall contrasted it to the melodramatic behaviour from other AI firms she said only served to “scare everyone”. Microsoft launched its own superintelligence team in October 2025 and, like Anthropic, is racing ahead with advanced AI development. Suleyman’s comments echo the company’s draft Humanist AI Code of Conduct, published this week, which champions “humanist superintelligence”. The code states that very advanced AI must stay within limits, and remain under human control. Comment now