Anthropic has disclosed in court filings that it cannot control or deactivate its AI models once deployed in Pentagon environments, lacking a 'kill switch' for such systems. This revelation comes amid ongoing legal disputes with the Pentagon, which has labeled Anthropic a supply chain risk due to concerns over the company's technology usage policies. These policies prohibit the use of Anthropic's AI in autonomous weapons and mass surveillance, which the Pentagon views as restrictive. The legal battle has resulted in conflicting court decisions: a Washington court upheld the Pentagon's risk designation, while a California court temporarily suspended it. Consequently, Anthropic is barred from new Pentagon contracts but can still serve other government agencies. Meanwhile, the Trump administration is advocating for the deployment of Anthropic's new AI model, Mythos, across federal agencies to bolster cybersecurity defenses, despite the Pentagon's security concerns. The next court hearing is set for May 19.