OpenAI’s head of ethics leaves less than a year after joining

That cynical view dominated, but the strongest version of it was more specific than "ethics is fake." People drew a distinction between classic policy-facing AI ethics and the kind of work labs now actually need. A frontier lab shipping powerful models cannot stop at essays about social harms. It needs testable criteria inside training and evaluation, because models will take stances and exploit loopholes whether the company planned for that or not. Several comments argued that this is the real transition underway. Ethics is being pulled out of the realm of principles and into the realm of implementation, where it starts to look like alignment targets, eval suites, guardrails, red-teaming, and product tradeoffs. Others answered that once ethics becomes implementation without independent power, it stops being ethics and becomes another way to rationalize what leadership already wanted. That tension shaped nearly every serious point. Some saw her exit as one more sign that OpenAI no longer wants internal friction after reorganizations, the Hugging Face hacking episode, and its steady move toward a more openly commercial posture. Others thought the role may never have had teeth in the first place, since big companies usually keep ethics functions only so employees, regulators, and investors can be told someone is handling the problem. A few commenters pushed back on overreading the move. Senior people leave large companies all the time, and quitting before an equity cliff can mean anything from a better offer to simple burnout. Still, the thread mostly landed on this: once a company’s strategy depends on moving fast with politically sensitive technology, "ethics" only survives if it can veto launches, shape evaluations, or impose costly constraints. If it cannot do those things, the role is PR with better vocabulary.