Hajra Naz
Anthropic has paused access to its latest AI systems after US authorities raised security concerns shortly after release. The decision affects Claude Fable 5 and Mythos 5, two of the company’s most advanced models.
The company said it was instructed to block use by foreign nationals. As a result, access to both models was disabled for all customers to stay compliant with the order.
Anthropic described Fable 5 as a highly advanced system. Internal messaging had previously labeled it “too powerful” for broad release.
Jailbreaking Concerns Trigger Government Action
US national security officials reportedly flagged a method that could bypass safety protections in Fable 5. This technique, known as jailbreaking, can remove software restrictions and expose hidden capabilities or sensitive functions.
Anthropic said the issue involved a limited number of known vulnerabilities. The company added that similar weaknesses can also be found in other publicly available AI models.
Before launch, Anthropic had highlighted safety safeguards and testing measures designed to reduce misuse. Early access was given to select organizations to evaluate risks in controlled environments.
Debate Over Safety vs Capability
Anthropic has defended its decision to test powerful models before public release. The company said Fable 5 showed stronger capabilities than previous versions, including potential use in cybersecurity-related tasks.
Critics, however, questioned the company’s earlier claim that the model was “too powerful to release,” calling it overstated marketing language.
Some researchers argue that restricting access may slow down independent safety research and limit collaboration between industry and governments.
Read More: Anthropic Explains Why Claude Fable 5 Safety Controls Were Not Visible
Global and Political Reactions
The European Union has also reacted to the development. Officials said the situation highlights the need for stronger technological independence from US and Asian AI providers.
Within the US, Anthropic has faced increasing political scrutiny. The Trump administration and senior defense officials previously labeled the company a “supply chain risk,” a classification rarely used for US-based firms.
That designation is linked to concerns about national security exposure and has become part of an ongoing legal dispute between Anthropic and US defense authorities.
Allegations of Foreign Access
A separate report from Semafor suggested that export restrictions may be linked to fears that Anthropic’s Mythos model could have been accessed by a China-related group. If confirmed, such access could raise concerns about model replication through distillation techniques, where smaller AI systems learn from larger ones.
The White House has not confirmed those claims. Anthropic has also disputed reports that jailbreaking vulnerabilities are widespread or unique to its systems.
Ongoing Legal and Security Pressure
Anthropic is currently involved in legal action with the Pentagon after being labeled a security risk. A US court has allowed government use of its tools to continue while the case proceeds.
At the same time, reports of possible breaches, including alleged unauthorized access through online communities, have added further pressure on the company’s most advanced models.
Read More: SpaceX Ignites New Debate Over Wall Street’s Top Tech Stocks: Mag 7 or MANGOS?
Industry Impact
The suspension highlights growing tension between rapid AI development and national security oversight.
As models become more capable, governments are increasing scrutiny over how they are released, tested, and accessed.
For Anthropic, the situation has turned its latest AI launch into a wider debate over safety controls, global access, and the limits of public deployment for frontier AI systems.
