More than 100 cybersecurity executives and researchers have urged the Trump administration to reverse restrictions placed on Anthropic’s most advanced AI systems, Fable 5 and Mythos 5. In a letter signed by experts from companies including Adobe and Nvidia, the group argued that preventing access to these models could weaken American cybersecurity capabilities rather than strengthen them. They contend that while the models are exceptionally capable at identifying and exploiting software vulnerabilities, they are not uniquely powerful compared with other advanced AI systems already available globally.
The dispute began after the U.S. government ordered Anthropic to suspend access to Fable 5 and Mythos 5 for foreign nationals, citing national security concerns. Officials reportedly feared that vulnerabilities or “jailbreak” techniques could allow the models to be used for offensive cyber operations or exploited by foreign intelligence services. To comply with the directive, Anthropic temporarily disabled the models for all users rather than attempting to selectively restrict access.
Anthropic has challenged the government's assessment, arguing that the alleged security issue was limited and not significantly different from risks present in other frontier AI models. The company stated that governments should have mechanisms to address genuinely dangerous deployments but that such actions should be transparent, technically grounded, and based on a clear process. The company has been engaged in ongoing discussions with U.S. officials in an effort to restore access to the models.
The controversy has evolved into a broader debate about AI governance, national security, and technological sovereignty. Critics of the restrictions warn that abruptly blocking access to advanced AI systems could undermine confidence among U.S. allies and encourage other countries to accelerate development of domestic alternatives. Supporters of the controls argue that frontier AI models should be treated as strategic technologies whose capabilities require careful oversight when national security risks emerge.
Beyond Anthropic itself, the case may become a landmark moment in AI policy. It represents one of the first instances in which the U.S. government has used export-control powers to restrict access to advanced AI models rather than hardware. The outcome could influence how future frontier AI systems are regulated, tested, and deployed worldwide, while shaping the balance between innovation, cybersecurity, and national security concerns.
