Key facts
- Nadella previously assumed AI alignment would automatically follow increased AI capability.
- He now believes alignment is a harder, different problem.
- Nadella previously warned about AI risks and advocated for external controls.
Microsoft CEO Satya Nadella has publicly stated that he has changed his perspective on artificial intelligence alignment. Previously, Nadella believed that as AI systems grew more capable, alignment with human values and goals would naturally occur. However, in a recent interview on the 20VC podcast, he expressed that this is not the case and that alignment presents a distinct and more challenging problem. Nadella's updated view aligns with concerns voiced by other prominent tech leaders, including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei, who have also cautioned about the potential threats posed by advanced AI. Earlier in October, Nadella had also emphasized the need for robust external controls and system design to manage AI models, likening them to insider threats.