The Consequences of Implementing AI in a Military and Defence Security Environment

在军事和国防安全环境中实施人工智能的后果

• 批准号: 1942837
• 金额: --
• 依托单位: Royal Holloway University of London
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2017
• 资助国家: 英国
• 起止时间: 2017 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-1942837
• 关键词: Consequences; Implementing; AI; Military; Defence

This research aims to answer the following research questions: AI as a weapon: how is AI developed, deployed, implemented and monitored in the military (and/or national security)? What controls are required to manage responsible and/or de-escalatory use of 'weaponised AI' capabilities? In order to approach these questions, the researcher is pursuing a grounded approach, developing theory based on strategy and policy documents as well as data gathered from the field through practitioner interviews, fieldwork and ethnographic observation. These methods involves attending a range of military and technical events looking at AI and defence, as well as a two month placement at the NATO cooperative cyber defence centre of excellence in Tallinn. By drawing insights on how AI technology is being implemented in practice in this environment, the researcher hopes to accurately map out security threats and analyse the relative success of mitigation for these threats (for example, through regional market regulation, military innovation policy, and the development, national and international norms. This research lies within EPSRC's remit alongside the goals of the information security PhD programme as it uses interdisciplinary methods to explain key technical security concerns in the national security environment. Placing emerging technology in the context of economic, social and political developments, this research will further the lively debate occurring around AI-enabled technical deployment in controversial environments, and how unforeseen negative consequences may be avoided.

本研究旨在回答以下研究问题：人工智能作为一种武器：人工智能如何在军事（和/或国家安全）中开发、部署、实施和监测？需要采取哪些控制措施来管理“武器化人工智能”能力的负责任和/或降级使用？为了解决这些问题，研究人员正在寻求一种接地的方法，发展理论的基础上的战略和政策文件，以及从外地收集的数据，通过从业人员访谈，实地考察和人种学观察。这些方法包括参加一系列关于人工智能和防御的军事和技术活动，以及在塔林的北约合作网络防御卓越中心进行为期两个月的实习。通过深入了解人工智能技术在这种环境中的实际应用，研究人员希望准确地绘制出安全威胁，并分析缓解这些威胁的相对成功（例如，通过区域市场监管，军事创新政策以及国家和国际规范的发展）。这项研究是在EPSRC的职权范围内，同时也是信息安全博士课程的目标，因为它使用跨学科的方法来解释国家安全环境中的关键技术安全问题。将新兴技术置于经济、社会和政治发展的背景下，这项研究将进一步推动围绕有争议环境中人工智能技术部署的激烈辩论，以及如何避免不可预见的负面后果。