Your Defense Code Is Already AI-Generated. Now What?

AI-generated code is already deeply embedded within the software supply chains of national defense systems, making policies that attempt to ban it unenforceable. Major commercial and open-source software, including operating systems and libraries used in defense applications, are increasingly built using AI coding assistants. This creates significant security risks, as research shows AI models can be trained to insert undetectable backdoors that evade standard safety checks. The focus for defense organizations must shift from debating adoption to building the verification infrastructure needed to manage the risks in systems they are already using.