The Trump administration has reportedly requested that OpenAI limit the release of its upcoming AI model, GPT-5.6, to a select group of government-approved partners before making it widely available. According to a source familiar with the situation, this move is driven by security concerns and represents the first instance where the U.S. government has sought to restrict the launch of an AI model before its official release. The White House’s Office of the National Cyber Director and the Office of Science and Technology Policy were involved in the discussions, aiming to establish a framework for testing and evaluating the security of such advanced models.
The decision follows internal communications within OpenAI, where CEO Sam Altman informed employees via a memo about the planned limited rollout of GPT-5.6. Altman expressed that this approach is not ideal for the company in the long run and emphasized the intention to collaborate with the government and other industry players to find a more sustainable method for future model releases. The administration had been engaged with OpenAI regarding the model's capabilities even before Anthropic removed access to its models due to a directive from the Commerce Department.
Behind the scenes, Altman met with Commerce Secretary Howard Lutnick to ensure that all relevant government entities had tested and approved the model. A source indicated that the government's intervention was motivated by the perceived power level of GPT-5.6, likening it to Anthropic's Mythos 5 model, which had previously faced restrictions. The concern stems from the potential risks associated with such powerful models falling into the wrong hands, whether through malicious intent or accidental misuse.
President Trump recently issued an executive order aimed at enhancing AI security, directing multiple agencies to create a voluntary testing protocol for AI companies prior to releasing new models. However, political debates over the extent of these requirements delayed the implementation of the order. As AI laboratories strive to stay ahead in the race for technological innovation, they face mounting pressure not only from competitors but also from increasing concerns about the implications of powerful AI models being accessible to potentially harmful actors.
OpenAI has confirmed that it will preview its latest GPT-5.6 models—Sol, Terra, and Luna—with a small group of trusted partners before a public release. Despite the administration's request, OpenAI maintains that such governmental involvement should not become standard practice. The company views the current limitation as a temporary measure, aiming to pave the way for broader accessibility in the near future while continuing to engage with the administration to refine cybersecurity frameworks and establish repeatable processes for future model launches.
In response to the administration's actions, there are growing concerns among experts and industry figures about the balance between ensuring AI safety and maintaining innovation momentum. Critics argue that without clear guidelines and standardized safety measures, the regulatory landscape could hinder progress in the field, potentially giving an advantage to international competitors like China. As OpenAI navigates these challenges, it remains committed to developing robust security features within its models, aiming to prevent misuse while promoting responsible AI development practices.
3 reports
AxiosIndependentCenterFactual 95Objective 9011 days agoThe Trump administration has requested that OpenAI limit the release of its upcoming GPT-5.6 model to a select group of government-approved partners before a broader launch, citing security concerns. This marks the first instance where the U.S. government has preemptively asked an American AI company to restrict a model's release. The request comes from the White House's Office of the National Cyber Director and Office of Science and Technology Policy, which aims to establish a framework for evaluating the security of new AI models. OpenAI CEO Sam Altman expressed his preference for a more open release strategy but acknowledged the need for collaboration with the government and industry stakeholders. The administration's interest appears to stem from concerns over the model's potential capabilities, similar to those of previously restricted models like Fable 5 and Mythos 5. The move aligns with a recent executive order aimed at improving AI security, though implementation has faced delays due to political disagreements.
Bias read (Center): The article presents information from multiple perspectives, including the administration's security concerns and OpenAI's response. It does not overtly favor one side over the other, nor does it exhibit strong ideological slant in its framing or sourcing. The focus remains on the technical and situ
Why these scores (Factual 95 · Objective 90): Accurately reports on the Trump administration's request to OpenAI regarding GPT-5.6, aligns with the primary source's claims about government intervention. Neutral tone with minimal bias.
The HillIndependentCenterFactual 85Objective 7510 days agoOpenAI has delayed the public release of its newest GPT-5.6 model, opting instead to first preview it with a select group of partners. This decision comes after consultations with the U.S. government, which has expressed concerns over potential risks associated with the advanced AI technology. The move suggests increased regulatory scrutiny and collaboration between tech companies and governmental authorities. While the exact nature of these concerns remains unspecified, the delay highlights growing tensions around AI development and oversight.
Bias read (Center): The article presents the situation as a response to government concerns without overtly favoring either the administration or OpenAI. It focuses on the collaborative aspect of the decision rather than taking a clear ideological stance.
Why these scores (Factual 85 · Objective 75): Factual accuracy aligns with the primary source document regarding OpenAI's restricted rollout of GPT-5.6. Objectivity is somewhat compromised by the article's focus on the Trump administration's role and implications, which introduces some bias.
TechCrunchIndependentCenterFactual 85Objective 7510 days agoOpenAI has restricted the rollout of its latest AI models, GPT-5.6, to a 'small group of trusted partners' at the request of the U.S. government. The models include Sol, Terra, and Luna, with Sol being the most powerful. This decision follows similar actions by the Trump administration against other AI firms like Anthropic, which was forced to remove its top model from public access. OpenAI expressed concern that such government interventions could set a precedent for ongoing restrictions on AI development and deployment. While the company complied with the current request, it emphasized that these measures should not become standard practice. OpenAI aims to work with the government to establish clearer guidelines for future AI model releases.
Bias read (Center): The article presents both the government's request and OpenAI's response without overtly favoring either side. It highlights concerns raised by OpenAI and a former White House advisor regarding the implications of government oversight on AI innovation, while also noting the administration's actions.
Why these scores (Factual 85 · Objective 75): Factually accurate in describing OpenAI's restriction of GPT-5.6 and linking it to broader government pressure on AI companies. Objectivity is slightly compromised by the inclusion of opinionated statements from Dean Ball about the implications of the restrictions.
★
Keep the news honest.
ObjectiveNews is reader-funded and ad-free — we show you the bias instead of hiding it. Support independent journalism for €5/month.
Become a Supporter
United States
