Chloé Bakalar joined OpenAI in August 2025 as the company’s first, and so far only, dedicated head of ethics. Eleven months later, in July 2026, she left again, with no public explanation. OpenAI has no plans to replace her. What sounds like a single personnel move fits into a broader pattern of departures from OpenAI’s safety and ethics ranks, and raises the question of what ethical oversight still means at a company currently preparing for what could be the largest tech IPO in history.
Key takeaways
- Chloé Bakalar, OpenAI’s sole dedicated ethicist, left the company in July 2026 after eleven months in the role.
- OpenAI is not filling the position, instead distributing ethical questions across multiple research teams.
- The same summer, Johannes Heidecke (head of Safety Systems) and Joshua Achiam (Chief Futurist) also left the company.
- Former robotics chief Caitlin Kalinowski had already left OpenAI in March 2026 over ethical concerns and moved to Anthropic.
- On August 11, 2026, CFO Brad Lightcap, an eight-year veteran of the company, also departed.
A role that existed exactly once
Bakalar joined OpenAI from Meta in August 2025, where she had previously helped shape AI ethics policy for Instagram and Facebook as chief ethicist. At OpenAI, multiple reports agree she was the only person at the company whose job was dedicated solely to the ethics of model development: how humans interact with AI systems, what limits should apply to chatbot personality design, and the ongoing debate over possible machine consciousness. On that last point, she had previously staked out a sober public position, describing language models as “prediction machines” whose capacity for sentience was “absolutely not” imminent. That stance stands in notable contrast to recent remarks from OpenAI CEO Sam Altman, who has publicly stated that humanity is already “in the singularity,” a contrast that makes the differing viewpoints inside the company visible.
Not an isolated case, but a pattern
Bakalar’s departure doesn’t stand alone. The same summer, two other executives with safety responsibilities left the company: Johannes Heidecke, who had led the Safety Systems team, and Joshua Achiam, OpenAI’s Chief Futurist. Back in March 2026, as kabel-salat.info reported at the time, robotics chief Caitlin Kalinowski left OpenAI on principle after failing to prevail in internal negotiations over surveillance and autonomous weapons systems for the Pentagon; she subsequently joined Anthropic. On August 11, 2026, another prominent departure followed: Brad Lightcap, who had spent eight years at OpenAI as CFO and most recently as head of special projects. No connection between these individual resignations has been established, but the cluster within just a few months is difficult to dismiss as coincidence.
“Ethics belongs to no one,” the company says
Asked about Bakalar’s departure, OpenAI responded with a line that has since been quoted nearly verbatim across multiple outlets: AI ethics “doesn’t live with one owner or team,” but is “deeply embedded into the model building process, driven by a number of teams across research.” For a company that created a dedicated ethics role and then eliminated it after eleven months, that reads like a post-hoc justification for a structural decision. Critics point out that distributed responsibility in practice often means no one is concretely accountable, an effect well documented in large organizations. Defenders of the move counter that a centralized ethics role in such a fast-growing company carries mostly symbolic weight anyway, if the actual decisions get made inside the research teams regardless. Critics raise a further structural point: a single ethics role at a company with several thousand employees was never going to review every model decision anyway, its real value lay in setting guardrails and, when necessary, publicly pushing back. That pushback function is exactly what disappears once no one is formally responsible for raising ethical concerns against the wishes of product teams.
Outlook
Whether dissolving the ethics role at OpenAI reflects a deliberate strategic shift or simply the outcome of several unrelated resignations can’t be settled from the outside. What stands out is the timing: the cluster of departures falls in a stretch where OpenAI is simultaneously announcing a voluntary AI evaluation framework and preparing for a possible IPO, precisely the moments when independent ethical oversight would matter most. Whether distributed responsibility actually works in practice will show mainly in whether OpenAI keeps talking publicly about ethical trade-offs in the coming months, or whether the topic quietly disappears along with the person who used to be responsible for it.