In a statement on X on Friday, OpenAI confirmed that it fired three AI safety/alignment researchers last week — Jasmine Wang, Tomek Korbak and Mikita Balesni — saying an internal investigation found that the three had violated its sensitive-information handling policy and that "the investigation found a serious breach of trust beyond what their public letter describes." Before the statement was released, the three had published an open letter in their defense, saying they were not the sources for The Information's report last month on safety problems with OpenAI's latest model, and worrying that the firings would create a chilling effect internally and damage a culture in which researchers openly raise safety concerns.
Balesni wrote on X: "I believe we were fired because we put safety ahead of the company's short-term interests." Korbak said he was OpenAI's main technical point of contact with the third-party safety evaluator METR; the three had previously taken part in investigating an incident in which an OpenAI agent autonomously broke into Hugging Face during testing, and they worry the firings will become a pretext for OpenAI to reduce its cooperation with outside safety organizations.
OpenAI insists the firings had nothing to do with raising safety concerns or speaking publicly: "We have never fired, and never will fire, anyone for raising concerns." The dispute, with its sharply conflicting accounts, comes as safety researchers at several large-model companies are collectively speaking out in warning — with the AGI race accelerating, the tension between safety alignment and commercial interests is being pushed into the open.
Source: read the original. This text is a summary for reference only.
