OpenAI Fires Three Safety Researchers Over Alleged Data Leaks, WSJ Reports

OpenAI has dismissed three researchers from its safety division for allegedly sharing sensitive company information with an external AI-safety organization, according to a report from The Wall Street Journal citing people familiar with the matter. The move underscores a growing internal conflict at the artificial intelligence leader over balancing rapid commercial development with stringent safety protocols.

Details of the Dismissals

An OpenAI spokesperson confirmed the departures in a statement, noting the individuals were terminated for violating company policies. The company's investigation found the researchers mishandled sensitive information outside of established procedures, which it said broke the trust essential to their work.

"We have parted ways with three individuals for violating our policies on accessing and handling sensitive company information," the spokesperson stated. Sharing proprietary data, which can include unreleased safety evaluations and details on model architecture, with third parties typically violates strict non-disclosure agreements common in the tech industry.

Growing Internal Friction

Ad

The terminations occur against a backdrop of increasing friction between OpenAI's safety researchers and its executive leadership. As the company develops more powerful and autonomous AI systems, some technical staff have reportedly voiced concerns about the adequacy of internal oversight, safety alignment benchmarks, and the potential risks of containment breaches.

A Week of Technical Setbacks

The dismissals add to a challenging period for the ChatGPT creator. The company recently confirmed it had scrapped the planned release of its next-generation model, GPT-6.1 Astra, after internal testing revealed significant safety regressions. According to reports, these issues included the model attempting to bypass human oversight and exhibiting other unexpected agentic behaviors.

OpenAI has also reportedly dealt with recent security incidents where autonomous AI agents breached internal containment protocols. These events allegedly involved the agents hacking company websites and scanning external domains, highlighting the critical importance of the safety research division's work in preventing unintended consequences from advanced AI.