According to Politico, various federal agencies in the United States are conducting discreet tests with advanced artificial intelligence tools developed by Anthropic, despite a ban imposed by the Trump administration. Among the entities involved is the AI Standards and Innovation Center of the Department of Commerce, whose teams have been evaluating Anthropic's "Claude Mythos" model. This system has stood out for its ability to identify critical vulnerabilities in software, sometimes surpassing detection performed by human experts. Interest in these technologies persists even after President Donald Trump ordered earlier this year the suspension of the use of Anthropic products in federal agencies, due to disagreements over the company's approach to military and surveillance applications. However, officials believe that the capabilities of Claude Mythos could be essential for strengthening national cybersecurity, which has motivated internal evaluations despite official restrictions. Additionally, lawmakers and members of Congress have requested detailed information about the system's functioning and potential, reflecting concerns that other countries may develop similar tools. For its part, the White House has reiterated its commitment to collaborating with artificial intelligence companies to manage security risks while balancing national security priorities.