OpenAI's next flagship AI model faces launch delay; White House demands safety checks

OpenAI's highly anticipated AI model, GPT 5.6, is facing unexpected delays as the company responds to requests from the Trump Administration for rigorous safety evaluations prior to its public introduction. As a result, the rollout will now be limited to a select group of partners, rather than a widespread release. CEO Sam Altman informed his team that access to the new model will be granted on a case-by-case basis during the preview phase. While OpenAI is working towards a broader launch within weeks, the government’s insistence on a cautious approach reflects growing concerns about the implications of advanced AI technologies. The administration's demands echo similar restrictions placed on other AI firms, including Anthropic, which faced limitations with its models, Claude Fable 5 and Mythos 5. OpenAI's collaboration with federal agencies, such as the Office of the National Cyber Director and the Office of Science and Technology Policy, indicates a proactive stance on ensuring the safety of AI advancements. In June, President Trump initiated an order directing major tech companies and AI developers to allow government evaluations of new models for safety and risk assessment. However, participation in these assessments remains voluntary, raising questions about regulatory compliance. Recent developments in AI have ignited cybersecurity fears, particularly the potential for these powerful models to be misused, leading to vulnerabilities within critical infrastructure and sensitive digital systems. Following the concerns surrounding Anthropic's Claude Mythos, which has now been restricted for broader use, the caution surrounding GPT 5.6 underscores the responsibility that comes with powerful AI technologies.