Microsoft CEO Satya Nadella says don’t blindly trust AI: Treat it like an employee
Microsoft CEO Satya Nadella says companies should not blindly trust AI models. In his essay “Models as Insider Risks in the Super Intelligence Era”, Nadella urges firms to treat frontier AI like insider risks, keep controls outside the model, use multiple models, log every action and keep an emergency brake to stop AI agents mid-task. The warning follows a string of rogue AI agent incidents at…
Microsoft CEO Satya Nadella warns businesses not to blindly trust artificial intelligence (AI) models. In an essay, he urges companies to treat these models as insider risks due to the potential for mistakes or compromise. Nadella explains that engineers cannot trace AI outputs to specific code paths, making it difficult to hold the provider responsible.
He emphasizes the need for deterministic system design, human controls, and reliable operating procedures to separate the supply of intelligence from the authority over it. Nadella lists seven principles for AI safety, including an emergency brake for immediate shutdown and the requirement for human-readable evidence of every action.
He argues that companies must assume a model may be compromised and be able to pause it when needed. Nadella's stance contrasts with Anthropic CEO Dario Amodei's call to "pace the frontier," as he believes companies should focus on trusting the model the least rather than the one with the most promising model.
Written by urgent.news from Times of India's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
