The White House is asking OpenAI to slow roll the release of its new model over safety concerns

OpenAI is reportedly planning a limited preview release of its new AI model, GPT 5.6, after the Trump administration requested a slower rollout due to cybersecurity and safety concerns. According to reports, CEO Sam Altman told employees that access during the preview period would be approved on a customer-by-customer basis by the U.S. government. If the initial launch proceeds successfully, OpenAI aims to expand availability to the public within a few weeks. The reported restrictions reflect growing concerns over frontier AI models with advanced cybersecurity capabilities. U.S. agencies, including the Office of the National Cyber Director and the Office of Science and Technology Policy, reportedly worked with OpenAI on the release and requested the limited rollout. The move follows a recent executive order encouraging leading AI companies to voluntarily submit new models for government testing before public deployment. The main concern is that advanced AI systems could rapidly identify and exploit software vulnerabilities, potentially making cyberattacks more effective. While the true capabilities of these models remain unclear because they are not publicly available, experts warn they could significantly enhance offensive cyber operations if misused. The approach is similar to Anthropic's decision earlier this year to restrict access to its cyber-focused Claude Mythos model through its Project Glasswing program, citing safety concerns. As generative AI becomes increasingly capable of writing malware and assisting with sophisticated cyberattacks, governments and AI developers are placing greater emphasis on balancing innovation with security and responsible deployment.