arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2510.08792cs.CY
收藏

面向国家安全的先进人工智能保障

Assurance of Frontier AI Built for National Security

Apollo Research(阿波罗研究)

Matteo Pistillo, Charlotte Stix

AI总结:

本备忘录针对人工智能错位带来的国家安全威胁,提出四项建议,通过利用现有测试评估流程和运营测试权限,开展谋划与控制评估,以强化模型可靠性与可治理性原则。

摘要:

本备忘录提出了四项建议,旨在加强人工智能模型可靠性和人工智能模型可治理性原则,因为国防部(DoW)、国家情报总监办公室(ODNI)、国家标准与技术研究院(NIST)和人工智能安全研究所(CAISI)正在根据《人工智能行动计划》完善人工智能保障框架。我们关注的焦点是错位这一开放科学问题及其对人工智能模型行为的影响。具体而言,错位和谋划能力可能是一个危险信号,表明人工智能模型的可靠性和可治理性不足。为了应对由错位引起的国家安全威胁,我们建议国防部(DoW)和情报界(IC)战略性地利用现有的测试与评估流程及其运营测试(OT)权限,通过一系列谋划与控制评估,为人工智能模型可靠性和可治理性原则提供面向未来的保障。

英文摘要:

This memorandum presents four recommendations aimed at strengthening the principles of AI model reliability and AI model governability, as DoW, ODNI, NIST, and CAISI refine AI assurance frameworks under the AI Action Plan. Our focus concerns the open scientific problem of misalignment and its implications on AI model behavior. Specifically, misalignment and scheming capabilities can be a red flag indicating AI model insufficient reliability and governability. To address the national security threats arising from misalignment, we recommend that DoW and the IC strategically leverage existing testing and evaluation pipelines and their OT authority to future proof the principles of AI model reliability and AI model governability through a suite of scheming and control evaluations.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑