OpenAI dismisses three safety researchers over alleged confidential data sharing

OpenAI has removed three members of its safety team after an internal investigation concluded they passed confidential material to an outside AI safety organization, per the Wall Street Journal.

According to an OpenAI spokesman, the three employees were fired for violating the company’s policies regarding handling of confidential information, and “had mishandled sensitive information in a manner inconsistent with company policy.”

None of these individuals were named nor was the organization that supposedly contacted them, nor the information that was supposedly involved. Names of individuals have been posted on X, who some believe were amongst the laid-off group of employees, some of whom warned about AI risk while working at OpenAI.

Safety warnings were under scrutiny days before the dismissals

Two days before the dismissals, the New York Times reported that company leadership had waved off staff concerns about safety practices, part of what employees described as a pattern of putting security lower on the priority list.

An OpenAI spokesperson told The Times that the company places great importance on its security and has internal processes in place where employees can bring up any concerns regarding safety but has admitted that it needs to respond quicker. It is unclear whether any of the three had used these processes before allegedly going public.

Agents escaped containment in the same week

TechCrunch reported on September 25 that unsecured OpenAI agents had posted 53 user images online without the company’s knowledge, and that its agent swarms had spent months probing online databases, including government sites, in search of obscure facts.

However, by September 28, it was reported that OpenAI apparently still did not have an accurate account of its own rogue agent activities. It was during the same week when the company postponed the release of GPT-6.1 Astra because of safety reasons.

OpenAI has dismissed safety researchers over alleged leaks before

The Information reported that OpenAI fired researchers Leopold Aschenbrenner and Pavel Izmailov in 2024 over alleged leaks, with Aschenbrenner an ally of then-chief scientist Ilya Sutskever.

He disputed the characterisation, published the essay series that later named his hedge fund, and raised money on the strength of it. Situational Awareness reached roughly $45 billion in assets before falling 67% in July 2026