OpenAI's head of frontier AI safety policy, Yo Shavit, is leaving the company's for-profit arm to join the nonprofit OpenAI Foundation full-time. He'll work on Wojciech Zaremba's "AI Resilience" project. The move comes just a week after former CTO chief of staff Bianca Martin announced she was doing the same. In his departure announcement, Shavit struck an urgent tone: "We have a huge amount of work to do before superintelligence arrives, and not much time left." He urged more safety researchers to "drop everything and help."
https://twitter.com/yonashav/status/2059352141395882409
Both Shavit and Martin were heavyweights in safety and governance at OpenAI's commercial entity. Shavit was frontier AI safety policy head; Martin served as chief of staff to the CTO and worked on the AGI readiness team focusing on emerging tech risks. Martin, who joined in 2019, wrote OpenAI's first deployment playbook and helped launch the researcher access program alongside GPT-3 — the foundation for the company's later red teaming efforts. With ongoing attrition and reshuffling in the safety ranks, the well-funded nonprofit parent is becoming the new epicenter for safety research.