OpenAI's head of ethics has left the company less than a year after joining, according to a report from the Financial Times. While the specific reasons for the departure haven't been fully disclosed, the exit raises serious questions about the state of AI safety governance at the world's most prominent AI lab.
A Pattern of Safety Departures
This isn't an isolated incident. OpenAI has seen a steady stream of safety and ethics researchers leave over the past two years, including:
- Members of the Superalignment team disbanded in 2024
- Multiple researchers who cited concerns about the pace of capability development outstripping safety work
- Board members who were removed during the November 2023 governance crisis and later reinstated with reduced oversight powers
Each departure follows a similar pattern: talented researchers join OpenAI to work on AI safety, become concerned about the gap between safety investments and capability investments, and eventually leave.
Why This Matters
The role of "head of ethics" at an AI lab isn't ceremonial — it's supposed to be the internal check on deployment decisions, data practices, and potential societal harms. When the person in that role leaves within a year, it suggests one of several possibilities:
The role lacks real authority. If the ethics team can't actually slow down or block deployments, the position becomes a fig leaf — there for appearances but without substance.
Internal pressure to ship overrides safety concerns. If the competitive dynamics with Anthropic, Google, and Meta create an environment where speed is paramount, ethics considerations get deprioritized.
Structural conflicts with leadership. If the CEO and board prioritize commercial goals over safety work, an ethics leader will inevitably clash with them.
Burnout from fighting uphill battles. Being the person who constantly says "wait, should we really do this?" in a ship-fast culture is exhausting.
The Broader Industry Context
OpenAI isn't alone in struggling with this tension. Every major AI lab faces the same structural problem: the people who build capabilities are rewarded with promotions, equity, and acclaim, while the people who flag risks are often seen as obstacles.
Anthropic was founded specifically to address this imbalance — it was created by former OpenAI researchers who felt safety wasn't being taken seriously enough. But even Anthropic has faced questions about whether its safety commitments can survive the pressure to compete.
Google's AI ethics history is well-documented, from Timnit Gebru's departure in 2020 to ongoing debates about how safety research interacts with product development.
What Needs to Change
The revolving door of AI ethics leadership suggests that the current model — hiring ethics researchers into for-profit AI labs and hoping they'll provide meaningful oversight — isn't working. Real alternatives include:
- Independent auditing bodies with regulatory authority, not just advisory power
- Mandatory safety assessments before model deployment, enforced by legislation
- Whistleblower protections for AI safety researchers who raise concerns
- Public transparency requirements — if an ethics team flags a risk, that flagging should be visible to regulators and the public
- Decoupling safety from profit — safety research funded independently of the companies building the models
The Bottom Line
When the person responsible for ethics at the most influential AI company in the world can't last a year in the role, it's not a personnel issue — it's a structural one. The system is working as designed; it's just that the design prioritizes capability over safety.
Until that changes, we can expect more departures, more warnings unheeded, and more risks realized. The question isn't whether the next ethics head will also leave — it's whether anyone will be paying attention when they do.
Based on reporting from the Financial Times, August 2026.
Top comments (0)