Ahead of Sam Altman's UN address, OpenAI proposes new ways to track AI misalignment risks
Sam Altman readies a global AI safety pitch for the UN.
TL;DR
- OpenAI released international AI safety standards as world leaders consider AI risk mitigation.
- The U.S. approach to AI risk coordination with China will be influenced by industry recommendations.
- OpenAI suggests leveraging AI safety institutes globally for standard-setting.
- The company proposes shared global measurements for classifying and responding to AI incidents and alignment issues.
- The goal is international standards for incidents in test settings and real-world applications.