The White House Just Invented AI Licensing Without Telling Congress

This is not regulation. This is a phone call.

Commerce Secretary Howard Lutnick allegedly called Sam Altman directly and told him not to launch GPT-5.6 without government approval. Not a rule. Not a law. Not even an executive order with teeth. A phone call. And OpenAI agreed.

If you are building frontier AI, pay close attention to what just happened - because this is how a de facto licensing regime is born.

The Fable Precedent

This request did not come out of nowhere. The Trump administration recently placed an export control order on Anthropic, forcing the company to pull its most advanced models - Mythos and Fable - off the market entirely. Washington and Wall Street panicked over their cybersecurity capabilities. The fear was that models this powerful could create unprecedented safety risks in the wrong hands.

OpenAI and the administration reportedly view GPT-5.6 as "on par" with Mythos. That is the threshold. If your model is scary enough, you do not get to release it. You get to ask permission.

Customer by Customer

In an internal memo, Altman told employees the government is now approving access "customer by customer." He made clear this is "not our preferred long term model," but OpenAI is playing along anyway - working with the government and "others in industry" to find a "more sustainable approach for future releases."

Translation: we know this is bad, but we are doing it because the alternative is being the next Anthropic.

A White House official told CNN they "continue to collaborate with frontier AI labs to develop shared approaches for addressing the challenges of scaling this technology." OpenAI declined to comment.

The Regulatory Vacuum

Here is what makes this genuinely dangerous. President Trump signed an executive order earlier this month asking AI companies with advanced models to voluntarily submit them for government review 30 days before release. But the framework for that review does not exist yet. No agency has been named. No criteria have been published. No appeals process. Nothing.

In the meantime, the confusion is total. The request to OpenAI came from the White House. The export control ban on Anthropic came from the Commerce Department. No one knows who is actually in charge of AI regulation, including the regulators.

There is no transparent, consistent framework for regulating AI right now - and that is exactly what makes ad hoc interventions so chilling.

Why This Matters for Every Builder

Brad Carson, head of the bipartisan pro-AI safety super PAC Public First, put it bluntly: "The Fable episode shows the need for clear regulations. Right now, you have an ad hoc, personalized, opaque, possibly lawless approach."

He is right. It is appropriate for the government to recall dangerous products, including AI models. But doing it through direct calls to CEOs, without transparency or basic fairness, sets a precedent that should terrify anyone building on the frontier.

Today it is OpenAI and Anthropic. Tomorrow it could be your startup. The threshold for "too capable" is undefined. The process for appeal is nonexistent. And the mechanism for enforcement is whatever agency feels like picking up the phone.

This is not safety regulation. This is improvisation at the highest stakes imaginable - and the entire industry just watched the biggest lab in the world agree to it.

Source: CNN Business