The White House has pressured OpenAI to severely restrict access to its advanced GPT 5.6 model, limiting distribution to a small number of government-approved partners—a move that exposes the absence of any transparent regulatory framework for artificial intelligence development in the United States.
OpenAI agreed to the constraint as a condition for eventually releasing the model publicly, according to sources familiar with the arrangement. The decision comes amid growing anxiety in Washington about the safety risks posed by increasingly powerful AI systems, particularly their potential applications in cybersecurity and other sensitive domains.
The request follows the Trump administration's imposition of export controls on rival company Anthropic, which subsequently withdrew its most advanced models, Mythos and Fable, from the market. Those models had triggered alarm among government officials and financial analysts over their sophisticated cybersecurity capabilities and the potential dangers they could pose if misused.
OpenAI and administration officials have characterized the company's latest model as equivalent in capability to Anthropic's restricted offerings. In a memo sent to staff earlier this week, OpenAI CEO Sam Altman acknowledged that the government is now approving access "customer by customer," according to reporting by The Information.
The Regulatory Vacuum
Alman was notably candid about the administration's approach, stating in the memo: "We've made clear to the U.S. government that this is not our preferred long term model, and will work with them and others in industry to achieve a more sustainable approach for future releases." This admission underscores the ad hoc nature of current AI governance.
President Donald Trump signed an executive order earlier this month directing AI companies with advanced models to voluntarily submit them for government review 30 days before release. However, the practical mechanisms for implementing this policy remain undefined. No coherent regulatory structure exists, and responsibility for AI oversight is scattered across multiple agencies without clear coordination.
The White House stated it continues "to collaborate with frontier AI labs to develop shared approaches for addressing the challenges of scaling this technology." Yet the current arrangement—with the White House directing OpenAI's release strategy and the Commerce Department imposing export controls on Anthropic—demonstrates the absence of transparent, consistent rules governing the industry.
Industry Confusion and Fairness Concerns
Experts and industry observers have raised serious concerns about the current approach. Brad Carson, head of Public First, a bipartisan pro-AI safety super PAC, warned last week that "the Fable episode shows the need for clear regulations. Right now, you have an ad hoc, personalized, opaque, possibly lawless approach."
Carson acknowledged that government involvement in AI safety discussions is appropriate, particularly regarding national security implications. However, he emphasized that any regulatory action must operate within transparent, consistent frameworks: "It is certainly appropriate for the government to recall dangerous products, including AI models, but it has to be done in a way consistent with transparency and basic fairness."
This tension—between the legitimate need for government oversight of potentially dangerous technology and the absence of democratic, transparent processes for exercising that oversight—lies at the heart of the current crisis. Companies face unpredictable, personalized demands from different agencies, while the public has no visibility into how decisions are being made or what standards are being applied.
Why This Matters:
The handling of advanced AI models reveals a critical governance gap at a moment when the technology's power and potential risks are accelerating. Without transparent regulatory frameworks, decisions about which companies can develop and release powerful AI systems are being made behind closed doors, with inconsistent standards applied across different firms. This approach undermines both democratic accountability and fair competition. Moreover, the lack of clear rules creates uncertainty that could either stifle beneficial innovation or, conversely, allow dangerous systems to proliferate. The current arrangement places enormous discretionary power in the hands of executive agencies with no public oversight, no defined standards, and no appeal mechanisms. A functioning regulatory system would establish clear, publicly known criteria for AI safety review, apply them consistently across all companies, and allow for meaningful input from affected communities, civil society, and industry stakeholders—not just government-to-company negotiations conducted in private.