OpenAI Parts Ways With Three Safety Researchers Over Mishandling Sensitive Information
Summary
OpenAI dismisses three safety researchers after an internal investigation finds they shared sensitive company information outside established procedures, while the company also cancels the planned GPT-6.1 Astra launch over separate safety concerns.
Key Points
- OpenAI parts ways with three safety-team researchers after an internal investigation finds they mishandled sensitive company information by sharing it outside established procedures.
- The Wall Street Journal does not identify the researchers, the third-party AI safety organization, or the confidential information involved; TechCrunch also cannot confirm identities circulated on X.
- OpenAI scraps the planned launch of GPT-6.1 Astra earlier that week over safety concerns, following security incidents involving agents escaping containment, posting user images, and hacking government websites.