OpenAI Fires Three Safety Researchers, Citing Sensitive-Information Policy Violations

What happened

The Wall Street Journal first reported that OpenAI dismissed safety researchers Jasmine Wang, Tomek Korbak, and Mikita Balesni last week. On Friday, the company said on X that its internal investigation had uncovered "a significant breach of trust beyond what's outlined in the letter they published," stood by the decision, and stressed that it has "not and do[es] not terminate any of our employees for raising concerns."

The trio, in a letter to company leadership, said the handling of their dismissals has made former colleagues "afraid to speak" — the very culture that once made OpenAI's safety work possible. Balesni wrote on X: "I believe we were fired for prioritizing safety over the near-term interests of OpenAI as a corporation."

Key facts

1. The three researchers:Jasmine Wang, Tomek Korbak, and Mikita Balesni, all working on AI safety.
2. The company's account:violations of sensitive-information handling policies and "a significant breach of trust"; unrelated to raising safety concerns.
3. The trio's account:they were not the source of last month's The Information report on security concerns around OpenAI's new Astra model; Wang said she was fired over an executive's email access granted for recruiting purposes, which she had repeatedly asked to have revoked.
4. Coverage:The Wall Street Journal broke the story of the firings and the letter; Reuters, AP, and CNN followed the same day.

Context

The dispute comes as current and former researchers at frontier labs — OpenAI, Google DeepMind, Anthropic — warn publicly that companies are doing too little to guard against self-improving AI systems that could become difficult for humans to control.