Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2601.05307 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915718071582720 |
|---|---|
| author | Gupta, Ritwik Reddie, Andrew W. |
| author_facet | Gupta, Ritwik Reddie, Andrew W. |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_05307 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | The LLM Mirage: Economic Interests and the Subversion of Weaponization Controls Gupta, Ritwik Reddie, Andrew W. Computers and Society 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. |
| title | The LLM Mirage: Economic Interests and the Subversion of Weaponization Controls |
| topic | Computers and Society |
| url | https://arxiv.org/abs/2601.05307 |