Anthropic Restricts Global Access to AI Models After U.S. Security Review

Anthropic has suspended access to its advanced AI models, Fable 5 and Mythos 5, for users outside the United States following new U.S. export restrictions, according to a Wall Street Journal report. The decision highlights growing government oversight of artificial intelligence technologies and their potential national security implications.

The restrictions reportedly came after discussions involving U.S. officials and Amazon.com, one of Anthropic’s largest investors and cloud infrastructure partners. According to the report, Amazon CEO Andy Jassy expressed concerns after researchers demonstrated that Fable 5 could be prompted to generate information that might assist cyberattacks. These findings triggered a review by White House officials and national security experts.

Following the review, U.S. authorities requested that Anthropic either address the identified vulnerabilities or limit access to the AI model. The administration ultimately concluded that restricting availability to foreign governments, businesses, and individuals was the most effective way to reduce potential security risks. President Donald Trump reportedly approved the measures despite concerns that tighter controls could slow innovation in the rapidly evolving artificial intelligence industry.

In response, Anthropic disabled international access to both Fable 5 and Mythos 5 while working to comply with the new requirements. The company has maintained that the issues identified were relatively basic and that similar capabilities can already be found in other publicly available AI systems.

The move represents a significant escalation in government scrutiny of frontier AI models. It also reflects broader efforts by U.S. policymakers to monitor advanced AI development and address concerns related to cybersecurity and national security.

Anthropic, which is reportedly preparing for a potential IPO later this year, may face additional challenges as a result of the restrictions. The company is best known for its Claude family of AI models, widely used by businesses and software developers worldwide.

Industry observers note that the restrictions could create opportunities for competitors such as OpenAI, as organizations seeking advanced AI solutions may explore alternative platforms while Anthropic works toward restoring broader access to its models.