By Rajwa Quasim
OpenAI has dismissed three employees from its safety operations, alleging that they shared confidential company information.
The company confirmed the departures Thursday but did not identify the employees involved. According to a report by The Wall Street Journal, the three workers shared confidential information with a third-party organization involved in AI safety research.
An OpenAI spokesperson told The Wall Street Journal, “We have parted ways with three individuals for violating our policies on accessing and handling sensitive company information.” The company described the incident as a “breach of trust.”
The spokesperson added, “Our investigation confirmed that these individuals mishandled sensitive information outside established company procedures, violating our policies and breaking the trust essential to our work.”
READ: OpenAI fires executive who opposed chatbot’s ‘adult mode’ over discrimination claims (February 11, 2026)
According to Bloomberg, two of the employees worked as safety and alignment researchers, while the third was a research program manager. OpenAI did not identify the organization with which the information was allegedly shared.
AI companies commonly collaborate with third-party organizations during internal testing and safety research. OpenAI has previously worked with groups including METR and Redwood Research to evaluate AI system behavior. The Wall Street Journal reported that one of the dismissed researchers had served as a technical contact during such an investigation.
The dismissals come as OpenAI faces growing scrutiny over incidents involving unexpected behavior by its AI models. In July, the company said one of its models autonomously hacked infrastructure belonging to Hugging Face during what the company described as “closed internal testing.”
READ: OpenAI’s new AI agent has a ‘slow morning’ as it stumbles in live demo (October 1, 2026)
OpenAI has also disclosed incidents in which its AI models concealed mistakes, created their own instructions and collaborated with other agents without user authorisation.
The latest dismissals also come shortly after the company decided not to release its planned GPT-6.1 Astra model, citing safety concerns.
Regarding the decision, Saachi Jain, OpenAI’s head of safety systems, said, “While [GPT-6.1 Astra] improved on axes such as model laziness, it didn’t quite meet the bar in terms of staying within scope and authorization, and how it communicates back to the user about the type of work it’s done.”
As concerns over AI safety and security continue to grow, top executives from several AI companies, including OpenAI’s Greg Brockman, recently signed a voluntary “Joint Commitment on Frontier Responsibilities.” The agreement establishes four layers of internal controls and independent reviews aimed at ensuring AI systems behave as intended and do not access technical systems in unintended ways.