Nadella alters views on AI alignment
The story so far
Microsoft CEO Nadella no longer believes that the stronger AI is, the safer it becomes, changing his views on the 'alignment' problem of AI. He previously thought that expanding model size could solve the alignment problem, but now believes it's more challenging. Nadella has warned about AI risks and advocated for strengthening external security controls, viewing AI models as internal security threats. He proposed establishing protective mechanisms for AI models whose behavior is difficult to fully predict through reliable, deterministic system design, human oversight, and sound operational procedures, and believes that for areas where existing standards are insufficient, industry standards need to be developed.
AI-written from the reports · Updated 23 days ago
Coverage timeline
Follow the coverage to see every side of the story.
- ITHome · AIMicrosoft CEO Nadella: No longer believes that AI is safer the more powerful it is.
Microsoft CEO Nadella has changed his views on the AI 'alignment' problem, no longer believing that more powerful AI systems are easier to align with human values. He previously thought that expanding model size could solve the alignment problem, but now believes it is more challenging. Nadella has previously warned of AI risks and advocated for strengthening external security controls, viewing AI models as internal security threats.