OpenAI’s head of ethics walked out after barely a year, a move reported by the Financial Times that has reignited doubts about the lab’s internal safety checks. The departure matters because the role is supposed to act as the company’s internal brake on risky deployments, and its sudden vacancy suggests that brake may be missing.

A pattern of exits

The resignation is not an isolated incident. Over the past two years, OpenAI has seen a string of safety-focused staff walk away. In 2024, several members of the Superalignment team left. Earlier, board members were removed during a governance scramble in 2023. The recurring theme is the same: researchers join to safeguard the technology, encounter friction, and exit.

What the role was meant to do

The head of ethics oversees deployment decisions, data-handling practices, and the internal vetting of potentially harmful capabilities. When the person occupying that seat departs quickly, it raises the question of whether the authority existed at all.

Possible reasons for the churn

  • Limited power to stop releases – If the role cannot compel a product team to halt a launch, the position becomes symbolic rather than functional.
  • Speed-over-safety pressure – Competition with other AI giants creates an environment where rapid feature delivery is prized, often at the expense of thorough risk assessment.
  • Commercial priorities eclipsing safety work – Leadership may favour revenue-generating projects, leaving safety teams understaffed and overburdened.
  • Burnout from constant friction – Repeated clashes between safety advocates and product engineers can wear down even the most committed researchers.

These factors echo a broader industry tension: labs reward those who ship new capabilities, while those who flag dangers are sometimes seen as obstacles.

Counter-point: internal safeguards may still exist

OpenAI has not publicly detailed the authority of its ethics head, and the company has not issued a statement on the resignation. It is possible that other internal mechanisms—such as cross-functional safety reviews or external advisory boards—continue to function. Without concrete evidence of their effectiveness, however, the vacancy still leaves a noticeable gap.

Why the departure matters beyond OpenAI

If a for-profit AI lab cannot retain an ethics lead for more than a year, the model of embedding safety experts inside commercial enterprises appears fragile. Anthropic, founded on the promise of safer AI, reports similar pressures, and even larger firms with extensive resources have documented friction between ethics units and product groups.

What could change the equation

  • Independent auditors with regulatory clout – External bodies that can enforce compliance, rather than relying on internal goodwill.
  • Legally mandated safety assessments – Rules that require documented risk analysis before any new model is released.
  • Stronger whistle-blower protections – Safeguards that let employees raise concerns without fear of retaliation.
  • Public disclosure of safety flags – Transparency that lets external experts and the public see what risks have been identified.
  • Funding safety research apart from profit motives – Grants or public financing that keep safety work insulated from product deadlines.

What to watch next

When an ethics leader cannot stay beyond a year, the underlying structure is called into question. Until the power dynamics between product ambition and safety responsibility are reshaped, the cycle of talent loss is likely to repeat. The next few months will show whether OpenAI—and the wider AI community—can break that pattern.