According to media reports, Sam Altman stated at an internal staff meeting that the U.S. government has requested the company to initially release GPT-5.6 to a select group of trusted partners before making it widely available. Altman emphasized that employees must comply with this directive, even if the company has reservations. This development follows less than two weeks after Anthropic was compelled to take down its advanced AI models under government pressure. The Trump administration is extending its intent to regulate cutting-edge AI models from Anthropic to OpenAI, signaling a further escalation in U.S. government intervention regarding top-tier artificial intelligence systems.

As reported by Bloomberg citing sources familiar with the matter, OpenAI CEO Sam Altman informed employees during an internal meeting this Wednesday that the U.S. government has mandated the company to limit the initial release of GPT-5.6 to a trusted partner group prior to a broader rollout. Altman stressed in the meeting that staff must cooperate with any requests from the Trump administration concerning safety and usage restrictions, even in cases of internal disagreement.

According to The Information, Altman communicated to employees in a Thursday memo that during the GPT-5.6 preview period, the government would approve access "on a customer-by-customer basis." Altman noted in the memo: "We have made it clear to the U.S. government that this is not our preferred long-term model, and we will work with the government and the industry to find a more sustainable approach for future product releases."

This move comes less than a fortnight after Anthropic was forced to globally remove access to its two most advanced AI products, Mythos 5 and Fable 5, following government pressure. OpenAI employees have reportedly expressed concerns that the government's actions against Anthropic could impact the broader deployment of ChatGPT technology.

Altman's Internal Directive: Compliance First, Despite Potential Disagreements

At the staff meeting, Altman clearly conveyed the government's stance, indicating that U.S. authorities are growing increasingly concerned about the capabilities of the latest generation of frontier models. He instructed employees to actively cooperate with government officials regarding safety and usage restrictions, framing this as an unavoidable compliance obligation.

Earlier this month, the U.S. government, citing national security concerns, required Anthropic to restrict the use of its models by foreign nationals both within and outside the United States. Anthropic subsequently disabled global access to its two most advanced AI products, Mythos 5 and Fable 5. Anthropic has since been engaged in negotiations with the government to seek restoration of access to these models.

Anthropic has publicly stated that the government's order was triggered by the discovery of a potential "jailbreak" vulnerability in its recently released Fable 5 model, which was already restricted from performing cybersecurity-related tasks. Anthropic contested the decision, arguing that "taking down a commercial model already deployed to hundreds of millions of users based on the discovery of a targeted potential jailbreak is a standard we believe is untenable. If applied consistently across the industry, this standard would effectively halt new model releases from all frontier model providers."

Regulatory Framework Remains Unclear as Industry Awaits Concrete Policies

Currently, AI labs rely on ad-hoc, self-developed processes when notifying the government about upcoming powerful model releases. Insiders indicate that it remains unclear which government department should be notified, when notifications should occur, how much detail should be shared, and how far in advance of a release such notifications are required.

Some industry observers are hopeful that an executive order signed by President Trump earlier this month will help clarify these issues. The order mandates that the Trump administration collaborate with AI companies to develop a voluntary framework within 60 days of signing. This framework is expected to include provisions allowing the government up to 30 days of advance access to so-called "frontier" models before their planned release.

Both OpenAI and Anthropic are racing to deploy more powerful new models across various domains, including programming and cybersecurity. As government intervention intensifies, finding a balance between technological advancement and compliance boundaries is becoming a central, inescapable challenge for the entire frontier AI industry.