Anthropic's Mythos AI model can find critical vulnerabilities in global financial, power, and government systems, triggering emergency responses from central banks and intelligence agencies worldwide.
Only the U.S. and Britain have received access to Mythos; the EU, China, Russia, and other nations are excluded, making it a major geopolitical leverage point for America.
Global leaders warn the model poses unprecedented cyber risks—the Bank of England governor called it capable of "cracking the whole cyber-risk world open," while Russia compared it to a nuclear bomb.
No international framework exists to manage Mythos-level AI risks; Anthropic estimates competitors will release similar models within 18 months, leaving little time for defensive fixes.
The scramble highlights how AI breakthroughs now function as geopolitical weapons rather than products, deepening technological divisions between nations.