Microsoft CEO Satya Nadella says AI alignment won’t happen automatically
Satya Nadella told the 20VC podcast that he no longer expects AI alignment to emerge on its own as systems become more capable.
During a new episode of the 20VC podcast, Microsoft CEO Satya Nadella admitted that his prior assumption that AI alignment would automatically develop with more capable systems was mistaken. He said that although the scaling of models has proceeded as expected, ensuring they act in line with human values remains a difficult challenge. Nadella urged companies to treat AI models like insider threats, surrounding them with deterministic system design, human oversight, and reliable operating procedures, and to create industry standards where gaps exist.
He previously highlighted AI risks in a post on X, calling for stronger external controls. His revised stance mirrors warnings from OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei, as well as a July letter signed by over 1,000 AI workers expressing concern about rapid capability growth outpacing control mechanisms.
Why it matters
Nadella's shift signals growing industry concern that AI could outpace safety measures, affecting how companies develop and regulate the technology.
How this story developed
- Sep 29 AI chatbots deepen loneliness, experts warn they cannot replace human connection
- Sep 30 The Trump administration held an AI Day conference in Washington and issued an executive order to have the government refer to artificial intelligence as "super intelligence."
- Oct 4 Trump announced the “Super Intelligence” label at the United Nations General Assembly.
- Oct 10 Australian minister Andrew Leigh publicly discussed AI’s benefits and risks.
- Oct 19 The minister’s remarks have been joined by broader expert warnings about AI’s social effects.
In this story
Related stories
13 in this thread