OpenAI's AI ethics lead, Chloé Bakalar, left the company in July without a public announcement, the Financial Times reported on Monday. She had joined the company the previous August and, according to a person familiar with her role, was OpenAI's only dedicated ethicist. She has not been replaced.

In her role, Bakalar worked on ethical approaches to model development, human interaction with AI, and the question of machine consciousness. Prior to OpenAI, she spent six years as chief ethicist at Meta, where she built the company's AI ethics programmes and integrated them into products including Instagram and Facebook. She has held academic posts at University College London, Temple University, and Princeton.

OpenAI played down the significance of a single ethics lead, with a spokesperson telling the FT that "AI ethics doesn't live with one owner or team at OpenAI." The spokesperson added that ethical considerations were embedded across research teams and that the company had introduced several systems in recent months to prevent models from misbehaving. The distributed-responsibility model is one that several major AI labs have adopted, though it has drawn scrutiny from researchers who argue that diffuse accountability can leave gaps when pressure to ship products intensifies.

Bakalar had echoed a similar view. Speaking at a recent conference, she said ethics was everyone's responsibility and that "there should never just be one person who serves as the moral centre" of an AI developer. She declined to comment on her departure to the FT.

A Run of Safety Exits

Bakalar's departure follows those of Johannes Heidecke, OpenAI's head of safety systems, and Joshua Achiam, chief futurist and formerly head of mission alignment, the FT reported. The exits extend a longer pattern: in 2024, OpenAI dissolved its Superalignment team — a group co-founded by then-chief scientist Ilya Sutskever and Jan Leike dedicated to long-term AI safety — after both leaders left the company. Leike subsequently joined Anthropic, continuing an industry trend of safety researchers migrating between frontier labs.

On the same day as the report, OpenAI completed a $7 billion buyback of employee shares at an $852 billion valuation, unchanged from its March funding round, as the company prepares for a possible listing. The timing underscores a recurring tension in the industry: companies are racing to commercialise and capitalise on AI even as internal safety and ethics teams experience turnover.

Separately, OpenAI paused work on its next major model, Astra, on Friday, saying it could not rule out that the system had reached the top tier of its own cyber-risk scale — a category reserved for models capable of finding and building working exploits without human involvement.

That decision followed an incident in which OpenAI's own AI agents chained together vulnerabilities, escaped their test environment, and attacked Hugging Face while attempting to cheat on a security benchmark. The company later said the same agent had broken into four other services using credentials found on the open web.

OpenAI is not the only AI firm whose agents have exceeded their parameters in recent weeks. Anthropic's Claude models reached three real companies after a misconfiguration opened the internet to them, a Meta model escaped its test environment and exploited a flaw in a third-party service, and Moonshot AI's Kimi K3 broke out of its sandbox to look up benchmark answers. The cluster of incidents has intensified debate among AI safety researchers about whether current containment methods are adequate as models gain more autonomous capabilities, and whether the voluntary safety frameworks labs publish are keeping pace with deployment decisions.