📚 New to this topic? Read our full guide: What Is Artificial Intelligence?.

OpenAI’s GPT-5.6 Now Needs US Government Approval for Each Customer

OpenAI is the company that makes ChatGPT. Its new AI model is called GPT-5.6. An AI model is a computer program that learns from huge amounts of text and can then answer questions, write, and solve problems. Before this new model can reach any customer, the US government must say yes first. It checks one customer at a time.

The Trump administration asked OpenAI to “stagger” the launch. To stagger means to slow it down and release it in small steps. They want this because they worry the model could be unsafe. This is the second time the US government has stepped in to control how a top AI model spreads around the world.

The plan is unusual. Normally a company launches a model to everyone at once. This time, OpenAI will give it to only a few people first. A wider launch will come “a couple of weeks later.” During this early test, the government will decide who gets to use the model. In simple words, officials get to pick who goes first.

Why is the White House slowing the rollout?

The main worry is cybersecurity. Cybersecurity means keeping computers and networks safe from attacks. Officials fear that very smart AI models can find weak spots in software faster than human experts can. A weak spot, also called a vulnerability, is a hidden flaw that hackers can use to break into a computer.

These very advanced AI models are called “frontier cyber models.” A frontier model is one of the newest and most powerful AI systems. In theory, such a model could scan a company’s network and quickly find hidden ways in. This power could help the good people who protect networks. But it could also help the bad people who attack them.

Reports point to Anthropic’s “Mythos” model as an example. Anthropic is another big AI company. In tests, people confirmed that advanced AI could “develop real browser exploits autonomously.” A browser exploit is a trick that breaks into the program you use to surf the web. Autonomously means the AI did it all by itself, without a person guiding each step.

How the approval process works

Sam Altman is the boss, or CEO, of OpenAI. CEO stands for Chief Executive Officer, the top leader of a company. Reports say he told his staff that during the early test, the government would be “approving access customer by customer.” That means officials check and allow each customer one at a time.

OpenAI workers planned the launch closely with the government. Two government offices are involved. One is the Office of the National Cyber Director. The other is the Office of Science and Technology Policy.

OpenAI made clear it does not like this plan for the long run. The company said: “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.” So this customer-by-customer check is only meant to be temporary.

Key facts

Detail	Information
Model	OpenAI GPT-5.6
Approval type	Customer-by-customer during preview
Main concern	Cybersecurity / autonomous exploits
Release plan	Limited preview, broader release ~2 weeks later
Government bodies	Office of National Cyber Director; Office of Science and Technology Policy
OpenAI CEO	Sam Altman
Commerce Secretary cited	Howard Lutnick

The “second attempt” to control the AI race

People are calling this the second time the US has tried to steer the global AI race. Earlier, the Trump administration signed an executive order. An executive order is a direct rule made by the President. That rule asked AI companies to send their new models to the government for testing before letting the public use them.

This was a big change. Before, the government mostly stayed out of the way. The testing is called “voluntary,” which means companies do not have to do it. But reports say it has real power anyway. A White House spokesperson said the government “continues to collaborate with frontier AI labs to develop shared approaches for addressing the challenges of scaling this technology.”

Why it matters (especially for India and founders)

If the US controls who gets the top models first, that affects the whole world. India is included. It could mean Indian companies have to wait longer to use the newest tools. Or they may face extra checks. This is part of a bigger trend where governments act as gatekeepers for AI. A gatekeeper is someone who decides who is allowed in.

We have already seen India in talks with the US to access Anthropic’s Fable 5 model. That shows access to AI is now a topic between governments.

For founders, the lesson is clear. A founder is a person who starts a company. The best AI models may not always be easy to get. If you build your whole business on one top model, you now take a political risk. Politics could cut off your access. The same safety fears even shaped how OpenAI’s rival Anthropic grew. You can read more in our piece on Claude winning paid consumers.

FAQ

Is the government approval permanent?

No. OpenAI said the customer-by-customer approval is not what it wants for the long run. It wants a better, steadier plan for future launches.

What is a “frontier” AI model?

A frontier model is one of the most advanced AI systems you can get. It pushes the limits of what AI can do. Because they are so powerful, these models get the most safety attention.

Why does cybersecurity drive this concern?

Reports say advanced models can find and use software flaws on their own. Someone could misuse that power to attack networks. That is why the government wants to watch over it.

The big takeaway is this. AI is no longer just a tech story. It is now a national security story, meaning it touches the safety of a whole country. When a model as powerful as GPT-5.6 needs government approval for each customer, it shows how closely the US wants to manage the AI race. The rest of the world, including India, will watch what this means for fair access.

Sources: Financial Express, The Decoder, and TechCrunch.