Anthropic Forced to Kill Access to Top-Tier AI Models After US Government Order

Anthropic announced it will abruptly shut down access to its most advanced AI systems, Fable 5 and Mythos 5, for all users after the US government issued an export control directive. The order, citing national security concerns, requires the company to suspend access by foreign nationals immediately - though officials provided no detailed explanation of the threat.

According to Anthropic, the government indicated a possible method exists to bypass safeguards on Fable 5, allowing it to be used for identifying software vulnerabilities. The company disputes the finding, calling it a "narrow, non-universal jailbreak" that shouldn't warrant recalling a product used by hundreds of millions.

This escalation marks a shift in US policy. Historically, export controls targeted AI chips and infrastructure, not the models themselves. The move also comes amid a strained relationship between Anthropic and the Trump administration. Earlier this year, Anthropic refused to let the US military deploy its models for domestic surveillance or fully autonomous weapons, prompting a supply chain blacklist set to take effect later this year.

The Pentagon's CIO, Kirsten Davies, publicly backed the decision, stating national security must come before "revenue cycles and pre-IPO valuation."

Anthropic, which confidentially filed for an IPO last month, had recently called for greater federal oversight of AI - including the power to block risky models. But it argues this directive violates principles of fair, evidence-based regulation. The company is working to restore access, saying it believes a misunderstanding occurred. Other Anthropic models remain unaffected.

Source: The Guardian