Nadella calls for an emergency brake on AI models
Microsoft’s CEO wants AI systems built with stronger controls, records and a human stop button as concern grows over losing control of models.
Microsoft CEO Satya Nadella has argued for tighter containment and auditability around advanced AI systems, saying they should be treated as potentially compromised from the outset. In a post on X, he urged a trust architecture that separates the model from the layer that manages its work, adds external safeguards, and leaves tamper-proof human-readable records of important actions. He also said an authorized person must always be able to pause or shut down a model while it is running, describing that capability as an emergency brake. The comments land amid growing industry concern about loss of control in AI systems and follow Anthropic CEO Dario Amodei’s call for more cautious development.
Why it matters
For companies building or deploying advanced AI, the point is that Microsoft’s chief is arguing for designs that assume failure and preserve a way to intervene. That shifts the focus from only checking outputs to making model actions auditable and stoppable while work is still under way, which is the control standard he says should be expected now.
Keep or strike?
Does this story matter, or is it hype? Mark it before you see what everyone else did.
Sources
- TechCrunch
- The Verge