AI-rewritten: This is a summary of an article from The Verge, rewritten by AI (Qwen, running locally) to make it easier to read. The facts come from the original article – read it for the full story.

---

Microsoft CEO Satya Nadella argues that highly advanced AI models should be treated as potentially compromised from the start. He urges the industry to stop viewing these systems as opaque "black boxes" and instead build transparent frameworks where models can be contained and observed. His plan includes ensuring all actions leave tamper-proof, human-readable evidence.

Nadella suggests implementing an emergency brake system so authorized personnel can pause or shut down any model mid-task. He notes that more advanced models will require specialized containment technologies that the industry must standardize. These ideas align with existing calls for timely incident disclosure and independent audits within the field.

Throughout his post, Nadella refers to AI as "super intelligence." While he emphasizes safety measures like verifiable data and containment, he does not provide specific dates or statistics regarding these risks. His focus remains on immediate technical controls rather than long-term predictions about future capabilities.