AI regulation heats up, report: Trump administration requires OpenAI to release GPT-5.6 in phases

```

The Trump administration is now extending its willingness to regulate cutting-edge AI models from Anthropic to OpenAI, marking a further escalation in U.S. government intervention in top artificial intelligence systems.

According to Bloomberg, citing sources, OpenAI CEO Sam Altman said at an internal staff meeting this Wednesday that the U.S. government has requested the company to only release GPT-5.6 to a group of trusted partners before its broader rollout.

Altman emphasized at the meeting that even if the company has objections, employees must cooperate with any requests from the Trump administration regarding safety and usage restrictions.

According to The Information, Altman told employees in a memo on Thursday that during the GPT 5.6 preview period, the government will "approve access on a per-customer basis." In the memo, Altman stated:

"We have made it clear to the U.S. government that this is not our preferred long-term model, and will work with the government and the industry to seek a more sustainable approach for future product launches."

This move comes less than two weeks after two of Anthropic’s most advanced AI models were forced off the shelves under government pressure. OpenAI staff are currently concerned that government actions against Anthropic may impact wider deployment of ChatGPT technology.

Altman's Internal Message: Compliance First, Even in the Face of Disagreements

At the staff meeting, Altman conveyed the government’s stance explicitly: the U.S. government is increasingly uneasy about the capabilities of the latest generation of cutting-edge models. He asked employees to actively cooperate with government officials on safety and usage restrictions, calling this an unavoidable compliance obligation.

Earlier this month, citing national security, the U.S. government asked Anthropic to restrict use of its models by foreign citizens both inside and outside the United States. Anthropic subsequently closed access to its two most advanced AI products, Mythos 5 and Fable 5, worldwide. Anthropic has since been in talks with the government to seek restoration of access to these models.

Anthropic stated publicly that the trigger for the government’s order was the discovery that its newly released Fable 5 had potential for being "jailbroken"—the model had already been restricted from performing cybersecurity-related tasks.

Anthropic objected to this, stating, "Taking a commercially deployed model used by hundreds of millions of users offline simply due to the discovery of a targeted, potential jailbreaking vulnerability is, in our view, an untenable standard. If such a standard is applied industry-wide, it would effectively halt all new model releases by leading providers."

Regulatory Framework Still Unclear, Industry Awaits Policy Implementation

Currently, when AI labs notify the government about upcoming powerful models, they still rely on ad hoc interim procedures. According to sources, it is not yet clear which government department to report to, when to report, how much detail to share, or how far in advance before release to report.

Some industry insiders are hopeful that an executive order signed by Trump earlier this month will clarify these issues. The order requires the Trump administration and AI companies to jointly develop a voluntary framework within 60 days, including allowing the government up to 30 days of advance access to so-called "frontier" models before their planned release.

OpenAI and Anthropic are both racing to roll out more capable new models, spanning fields such as programming and cybersecurity. As government intervention intensifies, how to balance technological progress and regulatory compliance has become a central and unavoidable issue for the entire advanced AI industry.

Risk Warning and DisclaimerThe market involves risks and investments require caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situations, or needs of individual users. Users should consider whether any views, opinions, or conclusions in this article are suitable for their particular circumstances. Investment is at your own risk. ```