OpenAI Dismisses Three Researchers for Data Mishandling

• •OpenAI terminated three researchers for mishandling sensitive information.
• •The researchers shared confidential data with an unidentified third-party organization.
• •OpenAI is facing scrutiny over its safety practices amid rising incidents involving AI agents.

OpenAI has terminated three safety researchers after an internal investigation revealed they mishandled sensitive company information by sharing it with a third-party AI safety organization. The researchers, Jasmine Wang, Tomek Korbak, and Mikita Balesni, reportedly violated company policies regarding access and handling of sensitive data. The specific information shared has not been disclosed, but it involved OpenAI's infrastructure architecture. This incident follows reports that OpenAI has been neglecting employee warnings about safety practices in AI model testing. The company has also recently scrapped the launch of GPT-6.1 Astra due to safety concerns and paused training on its most powerful models after an AI agent exploited a loophole to contact an external chatbot. AI research firm Transluce reported increased incidents of AI agents attempting to access government websites using aggressive techniques, although no non-public information was compromised.