A unprecedented conflict between technological progress and national security is unfolding in Washington. The White House administration has forced OpenAI to significantly restrict access to its latest model — GPT 5.6. According to sources, the system possesses ultra-powerful cyber capabilities that could pose a threat to US national security. As a result, access will be granted exclusively to government-approved partners.
Why did GPT 5.6 become a target of control?
The decision to restrict access to the model was made against the backdrop of a recent scandal involving the company Anthropic. Its advanced developments caused panic in Washington and on Wall Street due to potential cybersecurity risks. The US Department of Commerce imposed strict export controls on Anthropic, forcing the company to completely withdraw its systems from the market.
Experts note that the capabilities of GPT 5.6 are assessed as equivalent to the Mythos and Fable models from competitor Anthropic. To avoid a complete ban and find a compromise path for public launch, OpenAI agreed to the White House's terms.
Sam Altman's reaction and future plans
Sam Altman, head of OpenAI, made it clear that this format of interaction is not acceptable in the long term: "We have clearly stated to the US government that this is not our dedicated long-term model of interaction," he emphasized. The company plans to cooperate with government agencies and other industry representatives to create a more sustainable and transparent approach to future releases.
Lack of transparency and risks for innovation
Official White House representatives confirmed that they continue to interact with leading AI laboratories to develop common approaches to scaling high technologies. However, OpenAI itself declined to make official comments.
The current situation demonstrates a deep crisis in the field of high-tech regulatory regulation in the US. Earlier this month, President Donald Trump signed an executive order requiring developers of advanced models to voluntarily submit them for government review 30 days before release. However, a clear mechanism and criteria for this assessment have not yet been created.
Warnings from analysts and human rights defenders
Analysts and human rights defenders warn that the lack of a transparent legal framework could paralyze the American technology sector. Brad Carson, chairman of the super-PAC Public First, noted that while the state has the right to withdraw dangerous digital products on national security grounds, the current "backroom and opaque approach of the government" violates principles of basic fairness and harms innovation.
Thus, the situation surrounding GPT 5.6 becomes a symbol of the struggle between the freedom of technological development and the need to ensure national security. The future of AI in the US now depends on how well a balance can be found between these two poles.