According to TechCrunch, Lilian Weng, a leading safety researcher at OpenAI, announced her departure from the company. Weng, who served as Vice President of Research and Safety, revealed her decision on Friday, stating she is ready to explore new opportunities after seven years with the startup. Her last day will be November 15th, though she has not disclosed her future plans. Weng expressed pride in the achievements of the Safety Systems team and confidence in their continued success.
Weng's exit is part of a broader trend of departures from OpenAI, with several AI safety and policy researchers leaving the company over the past year. Some have criticized OpenAI for prioritizing commercial interests over AI safety. Weng joins other notable figures such as Ilya Sutskever and Jan Leike, who also left OpenAI this year to focus on AI safety elsewhere. Weng initially joined OpenAI in 2018, contributing to the robotics team that developed a robot hand capable of solving a Rubik's cube. As OpenAI shifted its focus to the GPT paradigm, Weng transitioned to building the applied AI research team in 2021 and later led the creation of a dedicated safety systems team in 2023.
Despite the growth of OpenAI's safety systems unit, which now includes over 80 scientists, researchers, and policy experts, concerns persist about the company's commitment to safety as it develops more powerful AI systems. Miles Brundage, a former policy researcher, left OpenAI in October, citing the dissolution of the AGI readiness team he advised. Additionally, former researcher Suchir Balaji expressed concerns about the potential societal harm of OpenAI's technology. OpenAI has stated that it is working on a transition plan to replace Weng and emphasized the importance of the Safety Systems team in ensuring the safety and reliability of its AI systems.
Other recent departures from OpenAI include CTO Mira Murati, Chief Research Officer Bob McGrew, and Research VP Barret Zoph. In August, prominent researcher Andrej Karpathy and co-founder John Schulman also announced their exits. Some of these individuals, including Leike and Schulman, have joined OpenAI competitor Anthropic, while others have pursued their own ventures.
