Anthropic AI Models Pulled: US Order Cites Security

US-based AI company Anthropic has disabled access to its two most advanced models, Fable 5 and Mythos 5. This affects all users outside the US and foreign nationals within the US. The move follows an export control directive from the US government, citing national security concerns. This is the first time a leading AI company has taken a publicly deployed model offline due to federal intervention. Commerce Secretary Howard Lutnick sent a letter to Anthropic, stating the models are subject to export controls. This means any export or transfer now requires an individual license, even for Anthropic's own foreign employees. Anthropic received the directive and disabled the models globally, as it couldn't selectively block foreign nationals in real time. The company noted the letter did not provide specific details on the national security concern. What's interesting is this shutdown comes just days after the models were launched. Fable 5 was a public release, while Mythos 5 was reserved for vetted cybersecurity defenders. An administration official indicated the Commerce Department acted after another company claimed it could "jailbreak" Mythos, raising alarms about national security risks. The administration had reportedly tried to get Anthropic to pause the release without success. Anthropic, however, pushed back, stating that a review of the "jailbreak" technique found only minor, previously known vulnerabilities. This situation highlights the growing tension between rapid AI development and government oversight for national security.