OpenAI announced Friday that it is restricting access to its powerful new artificial intelligence model, GPT-5.6 Sol, pending approval from the Trump administration—marking an unprecedented shift toward government control over which companies and customers can use cutting-edge AI technology.
The move represents a significant departure from how the tech industry has traditionally operated, placing federal officials in the position of deciding, on a case-by-case basis, which organizations gain access to advanced AI systems. OpenAI said the model would be accessible only to customers approved by the Trump administration, though the company has not disclosed the roughly 20 customers currently approved to use it.
OpenAI characterized the arrangement as temporary, describing it as a step on the "path to broader availability in the coming weeks." However, the company also stated: "We don't believe this kind of government access process should become the long-term default"—an implicit acknowledgment of concerns about the precedent being set.
The Cybersecurity Justification
The Trump administration's vetting framework stems from an executive order signed earlier in June that established a process for the federal government to review the national security risks of advanced AI systems for up to 30 days before public release. Officials have cited cybersecurity concerns as the rationale, particularly after Anthropic, OpenAI's chief rival, warned earlier this year that its Mythos model was adept at finding software flaws in ways that could be weaponized by malicious hackers.
OpenAI said its new Sol model "is better at helping people find and fix vulnerabilities" than it is at carrying out cyberattacks and does not cross the company's own risk threshold. The company acknowledged, however, that "there could be unforeseen risks, especially if the model is combined with other tools," citing this uncertainty as justification for stronger safeguards and a phased release.
Anthropic, facing similar pressures, announced Friday that the Trump administration had approved a limited release of its cybersecurity model, Mythos 5, two weeks after the U.S. Commerce Department effectively banned it. Anthropic said the government lifted restrictions, enabling the model to be "redeployed to a small group of cyber defenders and infrastructure providers." Commerce Secretary Howard Lutnick told Anthropic in a Friday letter that its work to address government concerns "yielded significant progress."
Democratic Accountability Questions
The arrangement has drawn criticism from those concerned about the lack of transparency and formal process governing these decisions. U.S. Rep. Lori Trahan, a Massachusetts Democrat and co-author of a bipartisan bill that would regulate AI, expressed concern in a statement: "The Trump administration is deciding company by company who gets access to the newest AI model. No law. No process. No oversight. Just appointees in Washington deciding who's in and who's out."
Stanford University cybersecurity expert Alex Stamos, chief product officer at AI security company Corridor and former chief security officer at Facebook parent Meta, challenged the factual basis for the restrictions. Speaking on a call with reporters earlier this week, Stamos said, "I just want to say that pretty much nobody in the cybersecurity industry believes that there's any factual basis for this action."
Stamos reviewed an analysis of research on Anthropic's Fable model conducted by Amazon, Anthropic's primary cloud computing backer, and said he "didn't find any risks that aren't present with other publicly available AI models, including those made in China." He concluded: "If the administration is honest about wanting the United States to beat China in this race, then this is about the dumbest thing they could possibly do."
Industry Negotiations and Power Dynamics
The restrictions have emerged from a series of negotiations between AI industry executives and Trump officials in recent weeks. OpenAI CEO Sam Altman spoke with U.S. Commerce Secretary Howard Lutnick about the model release Wednesday, part of these ongoing discussions.
Anthropic has also been engaged in talks with the administration, though CEO Dario Amodei has had a more contentious relationship with Trump officials. The Pentagon designated Anthropic as a national security risk for raising ethical and safety concerns about AI usage in war, and Trump himself ordered federal agencies to stop using Claude, Anthropic's chatbot. Anthropic responded with a lawsuit that is still working its way through federal courts.
Investor David Sacks, who co-leads Trump's council of technology and science advisers, blamed Anthropic for the heightened scrutiny, saying on a recent podcast: "Dario came to Washington a few months ago, back in April, and basically said that he had created a cyber weapon called Mythos" and "spiked the cortisol level, got everyone really worried." Sacks acknowledged, however, that "there was some truth to it in terms of the sense that this model had advanced cyber capabilities."
Anthropic said Friday it was "pleased" by the partial release of Mythos and will "continue to work with the government to expand access" and make Fable available again to general users.
Broader Implications for Public Ownership
The government's heightened AI oversight adds complexity to exploratory moves by OpenAI and Anthropic to take their companies public on Wall Street, following SpaceX's record-setting initial public offering on June 12. Trump has floated the possibility of the U.S. government owning a stake in leading AI companies, describing a concept where "pieces could be given to the American public, where the American public essentially becomes a partner with the companies."
The White House said Friday it continues to collaborate with frontier AI labs on addressing the challenges of scaling the fast-growing technology. The framework for government vetting, however, has not yet been fully developed, and the Trump administration described participation by AI developers as voluntary—even as the practical effect has been to restrict access.
Why This Matters:
This development raises fundamental questions about democratic accountability in technology governance. When unelected officials can unilaterally decide which private companies gain access to advanced technology—with no published criteria, no formal legal process, and no public oversight—it creates a system vulnerable to political favoritism and corporate capture. The fact that Anthropic, which raised safety concerns about AI in warfare, faces harsher restrictions than OpenAI, which has maintained closer relationships with the administration, suggests that compliance with political preferences may matter as much as actual cybersecurity risk. Additionally, the lack of transparency about the roughly 20 approved customers means the public cannot assess whether access decisions serve the public interest or private advantage. For workers and communities affected by AI deployment decisions, the absence of democratic process means their voices remain absent from decisions that could profoundly shape their futures. The proposed government ownership stake adds another layer of concern: without clear public interest frameworks, government ownership could entrench rather than democratize control over transformative technology.