On June 12, 2026, the US government told Anthropic to shut off access to two of its best AI models — Fable 5 and Mythos 5 — for anyone who isn't a US citizen. Anthropic confirmed this and news outlets like CNBC reported on the order.

The ban applies to foreign nationals themselves, not countries. That's an important difference. Normally, governments block things by country — say, "no one in Iran can access this." But this order says "only US citizens can access this," no matter where they live or work. Al Jazeera reported this detail too. It's like the rule follows the person instead of the passport stamp.

Fable 5 and Mythos 5 are Anthropic's most powerful AI models. When the US government restricts access to specific models like this, it's saying it considers them dangerous enough to control, the way it controls advanced computer chips and weapons technology.

Here's the problem: checking who someone is comes with real obstacles. If Anthropic wants to block access by country, it's easy — the company just looks at where you're logging in from. But to block access by nationality, Anthropic has to actually know who you are. That means asking users to prove they're US citizens or turning off access for anyone the company can't verify. Anthropic hasn't said exactly how it will do this yet.

For big companies with employees around the world, this creates an immediate headache. Imagine a company with headquarters in America and workers in Japan. Those Japanese workers might not be able to use Fable 5 anymore, even if they're working on a US company account. Lawyers and business managers at these companies will need to figure out fast if they're breaking the rules.

The US has been moving toward stricter AI rules over the past year. In the past, the government focused on controlling the specialized computer chips needed to build AI models. This directive is different: it blocks access to the finished models themselves. That's new territory.

What happens next is open. US-based AI companies like Anthropic now have to follow rules that their competitors in other countries don't face. Chinese and European AI companies can let foreign nationals use their models freely. That puts Anthropic at a disadvantage. Whether this is smart national security policy or an outdated approach to new technology is something Congress and affected companies will likely debate.

For now, the models are shut off. The rule is in place.