OpenAI is handing the keys to its next major model, ChatGPT 5.6, to the U.S. government before the public gets a look. The move forces a new level of collaboration—and oversight—between the world's leading AI developer and Washington.
A Mandated Preview Period
The decision appears to be influenced by several government bodies, including the Office of the National Cyber Director and the Office of Science and Technology Policy. The report also noted that Department of Commerce Secretary Howard Lutnick was involved in the discussions.
In the memo, Altman acknowledged this wasn't the company's ideal approach. "We've made clear to the US government that this is not our preferred long term model," he reportedly wrote.
Altman added that OpenAI will "work with them and others in industry to achieve a more sustainable approach for future releases."
This development follows a recent executive order from President Donald Trump, distinct from the Biden administration's earlier AI safety framework, which requested AI companies to participate in a federal review of their most powerful models before a public release. The government is expected to develop a standardized framework for these assessments.
How Voluntary is "Voluntary"?
The controlled release of ChatGPT 5.6 deepens the confusion surrounding the federal government's role in AI deployment. The move comes shortly after OpenAI's rival, Anthropic, disabled all access to two of its recent models following a federal directive.
That order, which aimed to block the tools from foreign nationals, did not specify the exact security concerns. The actions against both Anthropic and now OpenAI have raised questions across the industry about how the review process will function and just how voluntary it truly is.