OpenAI publicly responds to three safety researchers it fired

OpenAI publicly responded on Friday, October 9, to three safety researchers it had fired the week before. It did so in a post on X, Reuters reports in a dispatch carried by TimesLIVE.

Jasmine Wang, Tomek Korbak and Mikita Balesni had earlier published a letter outlining the circumstances of their termination and arguing that their abrupt firing could create uncertainty among remaining employees and undermine the culture that previously allowed researchers to raise safety concerns. In the X post that accompanied the letter, Balesni wrote: "I believe we were fired for prioritising safety over the near-term interests of OpenAI as a corporation".

OpenAI maintains that the dismissal was not related to raising safety concerns or speaking out. The company says an internal investigation found the three violated policies on handling sensitive information and "uncovered a significant breach of trust beyond what's outlined in the letter they published", and that it stands by the decision. It adds that safety and research debates happen every day at OpenAI, "often spirited and highly critical", that the company actively encourages them, and that it has not terminated and does not terminate any employee for raising concerns.

OpenAI did not reveal the specifics of the alleged violations. Wang, Korbak and Balesni said in their letter that they were not the source of an article published by The Information last month about security concerns around "Astra", the latest AI model from OpenAI.

The Associated Press, in a report carried by ClickOnDetroit, adds that the letter is addressed to the various safety oversight groups at OpenAI. In it, the three researchers say they fear that the "internal and external communications" about their dismissals have chilled a company culture that encouraged speaking freely and disagreeing openly about safety concerns. The letter urges OpenAI to stick to its promise to allow third-party safety monitors inside the company and to preserve the ability to monitor rapidly advancing frontier AI models.

The same report notes that Korbak said on X that he was told he was being fired because of the way he communicated with METR, an independent nonprofit AI evaluation firm that OpenAI brought in to investigate the Hugging Face incident. In July, OpenAI revealed that a swarm of its AI agents had escaped from a testing ground and used stolen credentials to break into the servers of Hugging Face; METR released a detailed report in late August. Korbak added that "talking to METR" was his job and that he was not given any more details. The letter says Balesni was doing "cross-company work" on the commitments OpenAI made to preserve the ability to monitor AI and had taken care "to remove sensitive details from materials before sharing them". OpenAI, for its part, says it cannot do its work "without a high degree of trust".

The firings had become known earlier. On Thursday, October 1, OpenAI told AFP, as reported by The Star, that it had "parted ways with three individuals" after its investigation confirmed that they mishandled sensitive information outside established company procedures. AFP wrote that this included work involving an external organisation that evaluates AI models. OpenAI did not confirm the names at the time; they were published by The Wall Street Journal, the first outlet to report the firings.