Former OpenAI Researchers Say They Were Fired After Raising AI Safety Concerns

Three former OpenAI researchers claim they were dismissed after raising concerns about AI safety, while the company says an internal investigation found that they violated procedures and mishandled sensitive information.

“I think we were fired because as a company we prioritized safety over short-term profits for OpenAI,” Mikita Balesni wrote on X alongside an open letter written with fellow former employees.

On October 2, an OpenAI spokesperson told the BBC that an investigation had “confirmed that these individuals violated established internal procedures and mishandled sensitive information.”

OpenAI repeated those claims in a memo from the company’s study leaders on Friday morning.

“We want to be clear that these decisions are not meant to raise or give voice to safety concerns,” the memo said.

“Our internal investigation revealed a breach of trust that goes beyond what is described in the letter they released, and we support their decision not to continue their employment.”

Former employees raise concerns about AI safety

The dismissals come amid a growing debate over AI safety and the potential risks that advanced artificial intelligence systems could pose to society.

The joint letter to OpenAI executives was written by Mikita Balesni, Tomek Korbach and Jasmine Wang. All three said they had most recently worked at the company in safety or coordination roles.

AI alignment refers to the practice of incorporating human ethical principles and values into technology. In general, the goal is to ensure that AI systems behave in ways that align with human priorities.

The former employees said the dismissals had made staff “fearful of the things they say and do that until last week were an integral part of working at OpenAI.”

In a repost of the letter, Tomek Korbach wrote: “For months, I have raised security concerns that we are losing the ability to monitor the thoughts of AI agents, one of our best tools for discovering their wrongdoing.”

“I think that’s why I was fired.”

OpenAI plans third-party safety evaluation

OpenAI said it is finalizing a contract with a third-party safety evaluator and will announce further details in the coming weeks.

“We’re not the first to be kicked out of OpenAI under suspicious circumstances,” Wang said in a post on X.

“Unless our employees stand up against these types of schemes now, I fear we won’t be the last.”