All AI is ‘compromised’ and has fewer rules than milk, Microsoft bosses say

Microsoft CEO Satya Nadella and President Brad Smith have issued stark warnings about artificial intelligence safety. Nadella wrote on X that all AI models should be treated as potentially compromised and dangerous, requiring built-in safeguards including the ability to pause or shut down a model mid-task. He compared this to an emergency brake and called for standardized containment technologies, especially for advanced superintelligence models, and said AI companies must adopt rigorous engineering approaches. Brad Smith told ITV News that AI is currently less regulated than washing machines or milk, arguing that urgent global safety rules are needed to build public trust. In a separate BBC interview, Smith reiterated the need for emergency brakes and kill switches to stop out-of-control AI systems. These calls follow similar warnings from Google DeepMind and Anthropic, gaining urgency after recent cyberattacks and incidents involving frontier AI models.

(Source：The Independent)