Imagine an AI technology so powerful that just three days after its launch, it is considered a threat to national security and banned. On June 9, 2026, Anthropic launched its most advanced models, Claude Fable 5 and Mythos 5, but these have now been banned worldwide.
Two Separate AI Models
Anthropic initially introduced this unlimited-capability AI of the Mythos class in two separate forms. The first was Fable 5, which was made available to the general public and came with strict safety guardrails. To handle risky queries, the company had also connected it with its previous model, Opus 4.8.
The second version, Mythos 5, had no safety restrictions and was given only to selected cybersecurity experts. The company claimed that this AI was the best tool so far for software engineering and scientific research.
The Real Reason Behind the Ban
The real reason behind this sudden action was the discovery of a “jailbreak” technique. According to a report by The Wall Street Journal, Amazon researchers found a way to bypass Fable 5’s safety guardrails using certain special prompts.
Through this technique, ordinary users could access the powerful and dangerous cyber capabilities of the Mythos model directly through Fable 5. The U.S. Department of Commerce feared that if this AI fell into the hands of hackers, it could be used to identify weaknesses in banks, government systems, and critical national infrastructure, leading to major cyberattacks.
Data, Evidence and Expert Opinion
The UK AI Safety Institute, in its rigorous assessment, found that the model was capable of easily breaching advanced security systems in 73% of cases. In addition, U.S. Defense Secretary Pete Hegseth described this AI tool as a “supply chain risk.” This is the first time that a product of an American company has been placed in such a threat category.
Based on these concerns, on June 12, 2026, the U.S. government used export-control rules to send an emergency directive to Anthropic. The order clearly stated that access to this AI must be immediately blocked for foreign nationals, including Indian users and foreign employees.
To comply with this strict legal order, Anthropic completely deactivated these models worldwide.
However, Anthropic disagrees with the administration’s decision. The company argues that the vulnerabilities identified through this jailbreak were very minor and that other AI models available in the market, such as GPT-5.5, can also detect such common weaknesses even without any jailbreak.
Consequences and Impact
This incident has created a major shockwave in the global technology world. Zoho founder Sridhar Vembu strongly reacted to the U.S. government’s strict action and described it as the “end of globalization.”
The U.S. government’s move proves that AI development is no longer just a commercial race; it has now become a highly sensitive weapon linked to national and global security.
Conclusion
Both of Anthropic’s powerful models have, for now, been put on hold for an indefinite period. In the future, it will be seen whether governments will fully control the future of AI or whether a new and safer path can be found between security and technological development.
