The U.S. government has imposed export control measures on Antropic's latest AI models, 'Claude Fable 5' and 'Mythos 5'. Antropic stated that the U.S. government issued a directive to halt access by foreign nationals based on national security authority, noting that this applies to foreign national employees within the United States as well. Consequently, Antropic has completely suspended access to both models to ensure compliance. The company explained that other Antropic models are not affected.

This measure is an example demonstrating that AI export controls could extend beyond semiconductors or computing infrastructure to include access to actual AI models. According to reports, the U.S. government expressed concern that Fable 5 and Mythos 5 could be exploited for purposes such as searching for security vulnerabilities, while Antropic countered that sufficient specific technical grounds were not provided.

Fable 5 is a high-performance model family from Antropic, known for its significantly enhanced coding and complex task execution capabilities. Domestic reports have also highlighted that Fable 5 demonstrated high performance in game development, creation of visualization tools, and long-duration tasks.

The controversy is expanding beyond a simple service suspension to include AI sovereignty and supply chain risks. The Canadian Prime Minister noted that this issue demonstrates how dangerous reliance on a specific U.S. AI provider can be. For companies, this reveals that relying on a single model or the AI infrastructure of a particular country can lead to the disruption of core operations due to a mere change in policy.

The core of this incident is 'who can use the most powerful AI, where, and under what conditions.' The AI competition is shifting beyond performance to issues of national security, export controls, and data access rights. Moving forward, it is highly likely that high-performance AI models will be treated not merely as software, but as strategic technologies.