Andrew W. Reddie

Ritwik Gupta, Andrew W. Reddie

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.

Ritwik Gupta, Leah Walker, Rodolfo Corona et al.

Current regulations on powerful AI capabilities are narrowly focused on "foundation" or "frontier" models. However, these terms are vague and inconsistently defined, leading to an unstable foundation for governance efforts. Critically, policy debates often fail to consider the data used with these models, despite the clear link between data and model performance. Even (relatively) "small" models that fall outside the typical definitions of foundation and frontier models can achieve equivalent outcomes when exposed to sufficiently specific datasets. In this work, we illustrate the importance of considering dataset size and content as essential factors in assessing the risks posed by models both today and in the future. More broadly, we emphasize the risk posed by over-regulating reactively and provide a path towards careful, quantitative evaluation of capabilities that can lead to a simplified regulatory environment.

Ritwik Gupta, Leah Walker, Andrew W. Reddie

U.S. export controls on semiconductors are widely known to be permeable, with the People's Republic of China (PRC) steadily creating state-of-the-art artificial intelligence (AI) models with exfiltrated chips. This paper presents the first concrete, public evidence of how leading PRC AI labs evade and circumvent U.S. export controls. We examine how Chinese companies, notably Tencent, are not only using chips that are restricted under U.S. export controls but are also finding ways to circumvent these regulations by using software and modeling techniques that maximize less capable hardware. Specifically, we argue that Tencent's ability to power its Hunyuan-Large model with non-export controlled NVIDIA H20s exemplifies broader gains in efficiency in machine learning that have eroded the moat that the United States initially built via its existing export controls. Finally, we examine the implications of this finding for the future of the United States' export control strategy.