AI Pulse · Frontier Labs Watch

OpenAI’s sole AI ethics lead, Chloé Bakalar, left the company last month after holding the role for less than a year. The position remains unfilled. This follows a pattern of departures from OpenAI’s safety and alignment teams over the past year, raising questions about how the company prioritizes internal oversight as it ships increasingly powerful models.

For enterprise buyers, this matters during vendor risk assessments. When you’re evaluating an AI provider, you want to know someone inside is asking hard questions about bias, misuse, and downstream risk. A vacant ethics seat, especially one that keeps turning over, signals that those conversations might not be happening with the weight they need. Compare this to Anthropic’s recent moves on watermarking and provenance, which signal a company building accountability into the product itself. The contrast is stark.

What this means for your procurement process

If you’re running diligence on OpenAI or any frontier lab, ask who owns ethics and safety decisions internally, how long they’ve been there, and whether they report directly to leadership or sit several layers down. Ask what happens when an ethics concern conflicts with a launch timeline. The answers, or lack of them, tell you how much risk you’re inheriting. This is exactly the kind of operational intelligence we build into tools like the Omni Command Centre, so your team can track vendor posture changes without manually refreshing news feeds.

The broader point is this: as AI moves from experimental to operational in your business, the stability and seriousness of your vendor’s internal guardrails become your problem. A revolving door in the ethics office is a data point. Treat it like one.