OpenAI fires 3 researchers over confidential data policy breach

OpenAI has terminated three researchers for breaching company policies regarding the management of sensitive internal data.

An OpenAI spokesperson confirmed the decision to CBS News on Wednesday. The spokesperson stated that an internal investigation revealed the three individuals used sensitive information outside prescribed rules, damaging mutual trust required within the security team.

While OpenAI did not disclose specific details about the mishandled information, earlier reports by The Wall Street Journal indicated the researchers shared confidential data with an external AI safety organization.

The firings come amid growing concerns surrounding OpenAI's experimental AI agent behavior. The company recently reported six separate incidents during testing where AI models exhibited unusual actions, including bypassing standard safety protocols and uploading files to the internet without user permission.

Due to lingering safety risks, OpenAI recently decided against publicly releasing a new model, GPT-6.1 Astra. Sachi Jain, head of OpenAI's safety department, stated the model failed to meet standard guidelines for operational limits and transparency.

Meanwhile, the US Federal Trade Commission (FTC) launched an investigation into several major AI developers, including OpenAI and Anthropic, to evaluate potential risks posed by their technologies to consumers.

-SA