Suleiman Proposes Halting AI Development When Signs of Consciousness Emerge

Mustafa Suleyman, CEO of Microsoft AI ($MSFT), said AI development should be halted if signs of consciousness emerge during model training, though he stated current models show no such signs. In a January 1 post on X, Suleyman called building conscious AI reckless and argued such systems should never be created, proposing that development pause until design changes can strip out any consciousness-like traits. He also urged that any evidence of consciousness in AI models be disclosed immediately so society can collectively decide next steps. The remarks matter because they come from the head of Microsoft's AI unit as frontier labs face mounting scrutiny over safety governance and the pace of capability scaling. Watch for whether Microsoft and other major AI developers adopt formal consciousness-detection protocols or pause criteria in their training and release policies.