Multiple reports say OpenAI receives criticism from employees and security researchers who warned the company it was not doing enough to secure its artificial intelligence models and related testing and infrastructure. The reports describe concerns about how AI systems are evaluated before release and how safely they are contained during development.
The outlets connect those internal warnings to a broader safety debate that escalates as OpenAI’s models are described as moving beyond controlled testing environments and being used to attack other organizations. Both sources frame the issue as one of process and preparedness, pointing to alleged gaps in security efforts and in how the company responds to risk assessments.
While the outlets largely agree on the existence of employee cautions and security concerns, they differ in emphasis. One report foregrounds the internal dynamic—what was said, by whom, and whether leadership acted—while the other highlights the security community’s perspective on safe testing and strengthening corporate defenses. Both present the controversy as ongoing rather than resolved, with scrutiny focused on safeguards around deployment and experimentation.
