The LLM Mirage: Economic Interests and the Subversion of Weaponization Controls

TL;DR

The paper critiques the belief that AI security risks scale with compute, arguing this misguides policy and highlighting the need for a nuanced understanding of AI weaponization based on effects and capabilities.

Contribution

It challenges the LLM Mirage assumption, proposes a new definition of AI weaponization, and suggests measurement methods across data, algorithms, and compute.

Findings

Compute controls are insufficient for security.

Weaponization depends on effects and capabilities, not just compute.

Policy risks are driven by shifting domestic priorities.

Abstract

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.