October 9, 2026 — OpenAI has publicly defended its decision to fire three safety researchers, saying the dismissals had nothing to do with raising safety concerns.
OpenAI said on Friday that it fired three researchers last week after an internal investigation found they violated the company’s policies on handling sensitive information. In a post on X, the company said its inquiry uncovered “a significant breach of trust beyond what’s outlined in the letter they published,” referring to the open letter the three former researchers — Jasmine Wang, Tomek Korbak, and Mikita Balesni — sent to OpenAI’s board earlier this week.
What the two sides are saying
The researchers’ letter argued their abrupt firing could chill internal safety debate, and Balesni said on X that he believes they were “fired for prioritizing safety over the near-term interests of OpenAI as a corporation.” OpenAI rejected that framing outright: “Safety and research debates happen every day at OpenAI, often spirited and highly critical. We actively encourage these discussions and consider them essential to making the right decisions,” the company said. “We have not and do not terminate any of our employees for raising concerns.”
The fired researchers said in their letter that they were not the source of a The Information article last month about security concerns around OpenAI’s latest model, GPT-6.1 Astra — the model OpenAI has since shelved after it failed internal safety tests. OpenAI did not disclose the specifics of the alleged violations.
Why it matters
Both sides agree on the stakes; they just describe the facts differently. Researchers at OpenAI, Google DeepMind, and Anthropic have warned publicly that companies are doing too little to guard against increasingly self-improving systems, and this episode — firings, letters, counter-statements — is exactly the kind of public fracture that makes researchers hesitate to flag the next problem. From a tested-tools perspective, the substance here is governance, not technology: if the people building the models don’t trust the process for raising alarms, the safety testing the industry keeps promising gets thinner over time. Watch whether any of the three lands at a competitor or a safety institute — that will tell you how much of this was a policy breach and how much was a culture clash.
Frequently asked questions
Why did OpenAI fire the three researchers?
OpenAI says an internal investigation found they violated policies on handling sensitive information — a “significant breach of trust” that went beyond what the researchers described in their public letter.
What did the researchers say about their firing?
They argued the dismissals were really about prioritizing safety over corporate interests, and warned the firings could silence remaining employees who might otherwise raise concerns.
Did the firings have anything to do with the GPT-6.1 Astra safety failures?
The researchers said they were not the source of The Information‘s reporting on Astra. OpenAI has not connected the firings to any specific model.
Is this related to the NYC Council hearing testimony?
Not directly, but the timing overlaps: OpenAI, Anthropic, Meta, and Google all faced questions about safety oversight under oath at the October 5 New York City Council AI hearing.
Sources: Reuters, OpenAI (statement on X)