A group of cybersecurity leaders has written to the Trump administration urging it to lift restrictions on Anthropic's artificial intelligence models, warning that current policy is undermining America's national security posture.
According to NDTV, the letter argues that removing access to top-tier cyber defense capabilities "without a good reason" is dangerous at a moment when America's adversaries are rapidly advancing their own AI programs.
The Financial Times went further, framing the restrictions as a strategic gift to China, specifically citing reduced access to Anthropic's model known as Mythos.
The push reflects a broader tension in Washington between export-control instincts — keeping powerful AI out of the wrong hands — and the operational needs of defense and intelligence communities that want to deploy the most capable tools available. Critics of the restrictions argue that over-regulating domestic use of frontier AI doesn't stop adversaries from building their own; it simply leaves American defenders with inferior options.
The signatories have not been fully detailed in available reports, but the framing of the letter centers on competitive risk: the argument is not that Anthropic's models are harmless, but that withholding them from U.S. cyber operators is a self-inflicted wound.
This matters because it signals growing friction between the national security community and the administration's approach to AI governance — a debate that will shape how the U.S. military and intelligence agencies are allowed to use commercial AI in the years ahead.