Fired OpenAI safety researchers deny misconduct and warn of chilling effect

In short: Three OpenAI safety researchers who were fired say the company’s misconduct claims are wrong, and they warn the firings may discourage safety work.

What happened

Jasmine Wang, Tomek Korbak, and Mikita Balesni, who worked on AI safety at OpenAI, published an open letter disputing why they were fired. They said OpenAI has claimed they mishandled sensitive information, but they deny breaking rules or working with outside groups beyond their job responsibilities.

OpenAI previously said the researchers were dismissed after allegedly sharing confidential company information with a third-party AI safety organization. In their letter, the researchers argue that safety teams often need to collaborate with outside experts, and that this kind of contact should be supported with clear internal processes.

The researchers also said their firings are having a “chilling effect,” meaning coworkers may feel afraid to speak up or ask hard questions. They compared this to a shift away from a culture where people could disagree openly about safety concerns.

OpenAI has not publicly responded directly to the letter, according to TechCrunch. However, the company shared an internal memo that praised the researchers’ contributions and said they were not fired for raising safety concerns.

Wang also posted more details on X, saying OpenAI told her she was fired for accessing an executive’s email inbox. She said she had been given that access for recruiting, tried to have it removed, and opened a sensitive email by mistake and reported it right away.

Why it matters

AI safety work depends on people being able to report risks early, even when it is uncomfortable. If employees think they could be punished for speaking up or for talking with outside reviewers (like inviting an inspector into a factory), problems can be harder to spot before they reach the public.

Source: TechCrunch AI