High-Risk AI Models Need Military-Grade Security

America's approach to securing artificial intelligence is dangerously inadequate, leaving the nation's most advanced AI research vulnerable to theft and sabotage. While the U.S. government champions a light-touch regulatory model to foster innovation, Chinese intelligence services are actively targeting American AI labs with sophisticated espionage. Proposals for simple transparency are insufficient because they treat these labs like commercial ventures instead of the strategic national assets they are. A tiered risk governance framework is needed, applying military-grade security to high-risk AI to prevent rivals like China from gaining a strategic monopoly on this transformative technology.