The LLM Mirage: Economic Interests and the Subversion of Weaponization Controls
For U.S. policymakers and national security analysts, the paper identifies a critical flaw in current AI security policy and offers a corrective framework.
The paper argues that U.S. AI security policy is misled by the 'LLM Mirage'—the belief that national security risks scale with compute used to train frontier language models—which miscalibrates strategy and destabilizes regulation. It proposes an intent-and-capability definition of AI weaponization and outlines measurement infrastructure across the AI Triad.
U.S. AI security policy is increasingly shaped by an $\textit{LLM Mirage}$, the belief that national security risks scale in proportion to the compute used to train frontier language models. That premise fails in two ways. It miscalibrates strategy because adversaries can obtain weaponizable capabilities with task-specific systems that use specialized data, algorithmic efficiency, and widely available hardware, while compute controls harden only a high-end perimeter. It also destabilizes regulation because, absent a settled definition of "AI weaponization," compute thresholds are easily renegotiated as domestic priorities shift, turning security policy into a proxy contest over industrial competitiveness. We analyze how the LLM Mirage took hold, propose an intent-and-capability definition of AI weaponization grounded in effects and international humanitarian law, and outline measurement infrastructure based on live benchmarks across the full AI Triad (data, algorithms, compute) for weaponization-relevant capabilities.