Image: David Dibert / pexels.com

A few weeks ago, the U.S. government made headlines around the world: Anthropic’s advanced AI model, “Claude Mythos 5,” was shut down practically overnight. The drastic measure was justified on the grounds of security concerns and potential risks to national security.

Now comes the surprising twist.

In a letter from the U.S. Department of Commerce, the ban was partially lifted. Selected companies and government agencies are once again permitted to use the AI model. But the case raises a much bigger question: How much control should a government have over modern artificial intelligence?

Why the AI Was Blocked in the First Place

The trigger was extraordinary. The U.S. government feared that the particularly powerful AI model could fall into the wrong hands or that security mechanisms could be circumvented. As a result, it issued an order at short notice to severely restrict access to the “Claude Mythos 5” models and, to some extent, “Claude Fable 5” as well.

The consequences were significant. It wasn't just customers outside the U.S. who were affected. Even the company's foreign employees were suddenly unable to access the systems. Anthropic had to completely disable access for many users on short notice in order to comply with the order.

The Surprising About-Face

Following intensive discussions between Anthropic and the U.S. government, partial approval has now been granted.

The Department of Commerce stated that sufficient security measures are now in place. As a result, more than 100 selected U.S. companies and government agencies are once again permitted to use the model. Organizations in the fields of cybersecurity and critical infrastructure are expected to benefit in particular. However, the restrictions remain in effect for the general public.

This shows that, even with the most advanced AI, political decisions are now just as important as technical innovations.

When Politics Determines Technology

This case highlights just how closely technology and politics are now intertwined.

Just a few years ago, companies largely decided for themselves when to release new software. Today, a government order can cause millions of users worldwide to lose access to a technology within a matter of hours.

Particularly in the field of artificial intelligence, many countries are increasingly concerned that powerful systems could be misused—for example, for cyberattacks, espionage, or other security-related purposes. As a result, governments are intervening more and more frequently.

For companies, however, this poses a significant risk. Even fully developed products can suddenly disappear from the market or be subject to restrictions on their use.

What does this mean for companies in Europe?

Even though the current case is taking place in the U.S., it affects companies worldwide.

Many European companies use American AI systems on a daily basis for programming, data analysis, or customer service. If access is restricted by political decisions, it could disrupt entire business processes.

The incident also highlights how dependent Europe currently is on a handful of major AI providers from the United States.

When a government agency in Washington issues an order, companies in Germany or Austria often feel the effects immediately.

This development is likely to further fuel the debate on digital independence in Europe.

The real challenge begins now

The key question is not whether countries should be allowed to regulate AI. Of course, dangerous technologies must be monitored.

The real challenge lies in striking a reasonable balance. Innovation must not be unnecessarily stifled. At the same time, security risks must be taken seriously. This tension will become increasingly important in the coming years, particularly in the field of artificial intelligence.

Companies need reliable rules. Developers need predictability. And users want to know whether the tools they rely on every day will even be available tomorrow.

The True Essence of the AI Decision

This case vividly illustrates just how much power governments now have to make decisions regarding digital technologies. Of course, a government must be able to intervene when there are genuine security threats. But sudden bans followed by a reversal achieve one thing above all else: uncertainty. Companies invest billions in new technologies and rely on clear rules. When political decisions are completely reversed within a matter of days, it undermines not only business confidence but also the drive for innovation. Anyone who wants to shape the future of AI needs transparent and reliable rules—not hasty, knee-jerk reactions.