A seemingly routine request from a Hill staffer can cause "dread" and initiate a month-long, multi-departmental review process within the Office of the Secretary of Defense (OSD) to ensure a unified and vetted response.
While veterans bring a crucial ground-truth perspective to policy, their experience is often narrow. Effective policymakers need to understand the broader bureaucratic machine, from the Pentagon to Congress, to see how the entire system functions.
China is far more willing to accept collateral damage and system failures. This tolerance could allow them to deploy less-than-perfect AI and robotic systems at scale, accepting high casualty rates (human and robotic) to achieve party objectives.
In Congress, requests from a member's personal office are often constituent-driven (e.g., district jobs). Requests from a Professional Staff Member (PSM) on a committee like HASC are issue-specific and tied to major legislation. This context is crucial for the DOD to interpret and respond appropriately.
The U.S. military's principle of using precise, minimal force requires developing highly sophisticated AI. In contrast, adversaries like Russia and China, who employ a "fire and forget" doctrine and tolerate civilian casualties, face a much lower technical bar for deploying autonomous systems.
Companies lobbying on policy should understand that no bill is implemented as originally drafted. It undergoes extensive horse-trading and revisions across committees, agencies, and industry interests, making the final outcome unpredictable and seldom perfect.
Standard AI evaluations use well-defined scenarios. Military operations are inherently dynamic and unpredictable. National security AI therefore requires a new evaluation paradigm focused on specific, tailored use cases and operational reliability under unforeseen circumstances.
AI's application in targeting is not monolithic. Tactically, it finds units (e.g., a tank). Operationally, it identifies key nodes to achieve objectives. Strategically, it discerns national pressure points to influence war outcomes, requiring vastly different data and models at each level.
Unlike the US model of a single "drone guy" per platoon, Ukraine has entire battalions focused on drone warfare. These units have frontline labs that debrief missions and iterate on drone technology within days, creating a dramatically faster innovation cycle.