OpenAI Faces Another Departure in Its Safety Leadership as VP Lilian Weng Steps Down

Key Takeaways

OpenAI’s VP of Research and Safety, Lilian Weng, is stepping down, marking another significant exit in the company’s safety leadership.

Weng’s departure follows similar exits, sparking concern over OpenAI's balance between pushing AI advancements and ensuring safety.

OpenAI is actively working to rebuild its safety team amid the recent wave of departures.

Lilian Weng, a central figure in OpenAI's safety research efforts and current VP of Research and Safety, has announced her resignation, effective November 15.

Her departure marks the end of a seven-year journey with OpenAI, starting in its robotics division. Here, she led a team to create the Rubik’s cube-solving robot hand—a project that showcased OpenAI’s innovation. In 2021, Weng transitioned to applied AI research, where she worked on high-impact AI applications. After the release of GPT-4, she established and led the company’s dedicated safety systems team, growing it to over 80 experts in research, science, and policy.

Weng’s decision to leave comes amid a series of high-profile exits from OpenAI, with many former researchers and executives openly questioning the company's focus. The latest departures include safety leaders like Ilya Sutskever and Jan Leike, who had pioneered the Superalignment team to guide superintelligent AI safely. Sutskever recently left to start his own AI venture, Safe Superintelligence (SSI), which quickly attracted $1 billion in funding, indicating a strong interest in safety-focused AI solutions.

Others, like Miles Brundage and Suchir Balaji, have raised concerns about OpenAI’s strategic direction, especially with the disbanding of teams like the AGI readiness team. Balaji shared his doubts online, expressing his skepticism around how OpenAI’s practices might affect public trust and legal standards for generative AI products.

With these high-profile departures, questions are being raised about OpenAI’s commitment to prioritizing safety, especially as the company dives deeper into commercial projects. OpenAI responded to Weng’s resignation with a statement of gratitude for her contributions, assuring that they remain fully committed to safety in AI. "Lilian’s work has been invaluable, and we’re grateful for her dedication to safety research,” an OpenAI spokesperson shared.

Beyond Weng, many key figures from OpenAI have moved to competitors like Anthropic or ventured into independent startups. Co-founders Durk Kingma and John Schulman, CTO Mira Murati, and chief research officer Bob McGrew are among those who've left, showing the growing tension in the AI industry as companies strive to balance innovation with responsible development.