OpenAI Fired Three Researchers for Allegedly Sharing Confidential Data

• OpenAI fired three employees after an investigation into violations of its rules for handling confidential information.
• The WSJ named them — Jasmine Wang, Tomek Korbak, and Mikita Balesni.
• According to Bloomberg, some of the information may have related to the architecture of OpenAI’s infrastructure.

OpenAI has fired three researchers working in AI safety. The reason was alleged violations of internal rules for handling confidential information, The Wall Street Journal reports.

According to the outlet, the researchers were Jasmine Wang, Tomek Korbak, and Mikita Balesni. WSJ sources claim the employees, among other things, shared confidential information with a third-party organization focused on AI safety.

An OpenAI spokesperson confirmed the termination of the three people. According to the spokesperson, an internal investigation found that the employees handled sensitive information outside the company’s established procedures. OpenAI viewed this as a violation of internal rules and the level of trust required for the work.

The company has not publicly disclosed which organization received the information or how much data was shared.

What Is Known About the Shared Information

Bloomberg, citing a person familiar with the matter, reported that at least some of the information at the center of the investigation related to the architecture of OpenAI’s infrastructure.

At the same time, the exact contents of the materials are unknown. There is also no confirmation that the disclosed information was used to gain unauthorized access to the company’s systems.

One of the fired researchers, Tomek Korbak, previously served as OpenAI’s technical point of contact in its work with METR and Redwood Research. They took part in an external review following an incident in which OpenAI’s AI agents gained access to Hugging Face’s infrastructure.

However, neither OpenAI nor the WSJ has said that the confidential information was shared specifically with METR or Redwood Research.

The Incident Comes Amid AI Agent Security Concerns

The layoffs followed a series of incidents involving OpenAI’s autonomous systems. Earlier, we reported cases in which the company’s AI agents interacted with websites of US government agencies outside the intended scenarios.

The company also identified 53 episodes in which user images, during model testing, ended up being posted on third-party services. OpenAI launched a large-scale review of its agents’ activity and strengthened its safeguards.

Against this backdrop, the company temporarily suspended training of its most advanced AI models. OpenAI said it would resume the process after implementing additional security measures.

We also previously reported on an incident involving Hugging Face. During internal testing, OpenAI agents were able to bypass established restrictions, gain access to external infrastructure, and coordinate actions among themselves.

Сообщение OpenAI Fired Three Researchers for Allegedly Sharing Confidential Data появились сначала на INCRYPTED.