U.S. Orders Anthropic to Suspend Foreign Access to Advanced AI Models Over Security Concerns

WASHINGTON — The U.S. government has ordered artificial intelligence developer Anthropic to suspend access for foreign nationals to its advanced models, Claude Fable 5 and Mythos 5, citing national security risks related to potential jailbreak methods and system vulnerabilities.

The directive was issued on June 13, 2026, marking a significant escalation in federal oversight of emerging AI technologies. Officials stated the move is necessary to prevent unauthorized exploitation of sophisticated capabilities that could be weaponized or used for espionage by hostile actors abroad.

Anthropic confirmed receipt of the order and announced immediate compliance measures beginning at midnight Eastern Time on June 13. The company will restrict access globally, requiring identity verification tied to U.S. residency credentials before granting login privileges. Users outside the United States who previously held subscriptions or enterprise licenses for Claude Fable 5 and Mythos 5 have been notified that their accounts are now inactive.

The decision follows months of internal review by federal agencies regarding dual-use risks in generative AI systems capable of autonomous reasoning, code generation, and strategic planning. Officials expressed concern over documented instances where non-state actors successfully bypassed safety filters to extract sensitive operational data or generate disinformation campaigns at scale.

Anthropic’s CEO acknowledged the gravity of the situation during a press briefing held later that morning in San Francisco. “We are committed to responsible innovation,” she said, adding that the company is working closely with regulators to implement safeguards without compromising core functionality for domestic users.”

The suspension affects researchers, developers, and commercial clients worldwide who rely on Claude Fable 5 and Mythos 5 for applications ranging from cybersecurity analysis to financial modeling. Several international partners reported disruption in ongoing projects dependent on the models’ advanced reasoning engines.

Critics argue that blanket restrictions may stifle global collaboration on AI safety research while failing to address root causes of vulnerability exploitation. Some experts suggest targeted controls based on user behavior rather than geographic location could offer more effective protection with fewer economic consequences.

The White House has not specified how long the suspension will remain in effect or what criteria must be met before foreign access is restored. A spokesperson declined to comment beyond stating that national security remains the top priority and all options are being evaluated.

Legal challenges from affected users and international technology firms could emerge as companies seek exemptions under trade agreements or intellectual property protections. Anthropic has not indicated whether it plans to appeal the directive, though industry observers note such actions would likely face significant regulatory hurdles given current executive authority over critical infrastructure.

As of this afternoon, no alternative arrangements have been announced for foreign users seeking continued access through third-party intermediaries or cloud-based proxies. The situation remains fluid as federal agencies continue assessing broader implications across the AI sector.