The Trump Administration has requested OpenAI limit its upcoming GPT 5.6 model release to a small number of government-approved partners, citing advanced cybersecurity capabilities that officials fear could pose unprecedented safety risks.

OpenAI agreed to the restriction as a conditional pathway toward eventual public launch, exposing a fundamental problem that the US lacks any transparent regulatory framework for AI model oversight.

According to the memo by OpenAI CEO Sam Altman shared with employees on Thursday, the government is now granting approval to each company "customer by customer".

According to the White House, the GPT 5.6 model of OpenAI has capabilities similar to Anthropic's Mythos model, which has been withdrawn from the commerce department in mid-June due to concerns regarding cybersecurity.

According to The Information, Altman conveyed OpenAI's displeasure over the process: "We've made clear to the US government that this is not our preferred long-term model and will work with them and others in industry to achieve a more sustainable approach for future releases."

An executive order was signed by President Donald Trump last month requiring AI companies to provide advanced models for a period of 30 days for governmental review prior to the model being released. However, the mechanism has yet to be set up.

Public First Head Brad Carson, a bipartisan pro-AI safety super PAC, told CNN: "Right now, you have an ad hoc, personalised, opaque, possibly lawless approach. It is certainly appropriate for the government to recall dangerous products, including AI models, but it has to be done in a way consistent with transparency and basic fairness."

OpenAI declined to comment beyond Altman's memo. A White House official said the administration continues "collaborating with frontier AI labs to develop shared approaches for addressing the challenges of scaling this technology".