Professional Documents
Culture Documents
Artificial Intelligence in Armed Forces 8259b5d9 Merged Public
Artificial Intelligence in Armed Forces 8259b5d9 Merged Public
Introduction Introduction
Today is the age of the Fourth Industrial Revolution. The current period 今天是第四次工业革命的时代。 本期
of rapid, simultaneous and systemic transformations driven by advances 由进步驱动的快速、同时和系统的转变
in science is reshaping industries, blurring geographical boundaries, 科学正在重塑行业,模糊地理界限,
challenging existing regulatory frameworks and even redefining what it 挑战现有的监管框架,甚至重新定义它
means to be human. Artificial Intelligence (AI) is the software engine 意味着成为人类。 人工智能(AI)是软件引擎
that drives the Fourth Industrial Revolution. AI is creating targeted 这推动了第四次工业革命。 人工智能正在创造有针对性的
marketing, safer travel through self-driving cars, smarter weapons and 营销,通过自动驾驶汽车更安全的旅行,更智能的武器和
new efficiencies in manufacturing processes, supply chain management 制造流程、供应链管理的新效率
and agricultural production. It holds the promise of solving some of the 和农业生产。 它有解决一些问题的承诺。
most pressing issues facing society, but also presents challenges such as 社会面临的最紧迫的问题,但也提出了挑战,如
inscrutable “black box” algorithms, unethical use of data and potential 高深莫测的"黑盒子"算法,不道德地使用数据和潜力
job displacement. 工作位移。
Artificial intelligence comprises machine programmes that can 人工智能包括机器程序,可以
teach themselves by harnessing High Power Computing(HPC) and big 通过利用高功率计算(HPC)和大
data, and eventually mimic how the human brain thinks, supports and 数据,并最终模仿人类大脑如何思考,支持和
enables nearly every sector of the modern economy. Corporations and 使现代经济的几乎每一个部门都成为可能。 法团及
governments are fiercely competing because whoever is the frontrunner 各国政府正在激烈竞争,因为谁是领跑者
in AI research and applications will accrue the highest profits in this fast 在人工智能的研究和应用将在这个快速的过程中获得最高的利润
Major General P K Mallick VSM (Retd) was a Senior Directing Staff (SDS) at the National P K Mallick vsm少将(Retd)是国家的高级导演人员(SDS)
Defence College, New Delhi. 国防学院,新德里。
growing market and gain a military technological edge. AI itself will 不断增长的市场,并获得军事技术优势。 AI本身会
not manifest just as a weapon. It is an enabler that can support a broad 不仅仅是作为武器表现出来。 它是一个使能者,可以支持广泛的
1
spectrum of technologies. 技术的频谱。 1
These technologies are starting to have a transformative effect on 这些技术开始对......产生变革性影响。
defence capability. AI will have digital, physical and political security 防御能力。 人工智能将拥有数字,物理和政治安全
implications, expanding existing threats, introducing new threats and 影响,扩大现有威胁,引入新的威胁和
changing the character of threats and of war. These changes could include 改变威胁和战争的性质。 这些变化可能包括
the automation of social engineering attacks, vulnerability discovery, 社会工程攻击的自动化,漏洞发现,
influence campaigns, terrorist repurposing of commercial AI systems, 影响运动,恐怖分子重新利用商业AI系统,
increased scale of attacks, and manipulation of information availability. 攻击规模的增加和信息可用性的操纵。
This has caused countries around the world to become increasingly aware 这使世界各国日益意识到
of the impact of AI. There has been tremendous activity concerning AI AI的影响。 关于人工智能的活动非常活跃
policy positions and the development of an AI ecosystem in different 不同领域的政策立场和人工智能生态系统的发展
countries over the last 18 to 24 months. To cite a few examples, the 过去18至24个月的国家。 举几个例子,
US published its AI report in December 2016; France published its AI 美国于2016年12月发布了人工智能报告;法国发布了人工智能报告
strategy in January 2017 followed by a detailed policy document in March 2017年1月的战略,随后在3月发布了详细的政策文件
2018; Japan released a document in March 2017; China published its AI 2018年;日本于2017年3月发布文件;中国发布了AI
strategy in July 2017; India published a Task Force report in August 2017年7月的战略;印度8月发表了一份工作队报告
2017, and the UK released its industrial strategy in November 2017. 2017年,英国于2017年11月发布了产业战略。
Global investments in artificial intelligence for economic and national 面向经济和国家的人工智能全球投资
security purposes are increasingly being described as an arms race. The 安全目的越来越多地被描述为军备竞赛。 该
character of AI technology, like robotics, makes many countries well- 人工智能技术的特点,就像机器人一样,使许多国家都很好-
positioned to design and deploy it for military purposes. In Southeast 定位为军事目的设计和部署它。 在东南部
Asia, Singapore is on the cutting edge of AI investments (both military 在亚洲,新加坡处于人工智能投资的前沿(包括军事投资)
and non-military). In the military domain, South Korea has developed 和非军事)。 在军事领域,韩国发展
the SGR-A1, a semi-autonomous weapon system designed to protect the SGR-A1是一种半自主武器系统,旨在保护
demilitarised zone from attack by North Korea and others. 非军事区不受朝鲜和其他国家的攻击。
This explains that AI is increasingly becoming a key component 这解释了人工智能正日益成为一个关键组成部分
2
of national security. As India’s 2017 Task Force report states: “AI can 国家安全。 正如印度2017年特别工作组报告所述:2"AI可以
be a force-multiplier for several national security missions. Potential 成为几个国家安全任务的力量倍增器。 潜力
applications include Autonomous Underwater Vehicles (AUVs) and 应用包括自主水下航行器(AUVs)和
Unmanned Combat Aerial Vehicles (UCAVs) for underwater and aerial 水下和空中无人作战飞行器
defence operations as well as coastal and border surveillance; robot mules 国防行动以及沿海和边境监视;机器人骡子
for unburdening soldiers; robots for counter-IED and counter-terrorism 用于减轻负担的士兵;用于反恐和反恐的机器人
operations and close-in-protection systems against cruise missiles and 针对巡航导弹和
3
similar aerial threats […]”. In view of this, national security imperatives 类似的空中威胁[...]"。 3有鉴于此国家安全当务之急
require that technology-based force multipliers be developed. In doing 要求开发基于技术的力乘数。 在做
so, some areas where AI-based systems could be usefully deployed are: 因此,可以有效部署基于人工智能的系统的一些领域是:
autonomous surveillance and combat systems, adaptive communications 自主监视和作战系统,自适应通信
systems, AI-based cyber attack mitigation and counter-attack systems 系统、基于人工智能的网络攻击缓解和反击系统
and, multi-sensor data fusion-based decision-making systems. 以及基于多传感器数据融合的决策系统。
about the nature of war.8 Former US Deputy Secretary of Defence 关于战争的性质。 8前美国国防部副部长
Robert O. Work said, “Rapid advances in artificial intelligence and the 罗伯特*O*工作说:"人工智能和
vastly improved autonomous systems and operations they will enable are 大大改进的自治系统和操作,他们将使是
pointing towards new and more novel war-fighting applications involving 指向新的和更新颖的战争应用,涉及
9
human-machine collaboration and combat teaming”. The role of 人机协作和战斗组队"。 9的作用
educated humans will begin to concentrate on the higher cognitive tasks of 受过教育的人类将开始专注于更高的认知任务
processes such as mission analysis, operational planning, and assessments. 任务分析、业务规划和评估等过程。
More specifically, in relation to future conflict, as ambiguity will increase 更具体地说,关于未来的冲突,由于模糊性会增加
despite interconnectedness, the velocity and scale of activity will make 尽管相互联系,但活动的速度和规模将使
it difficult to discern the important from the unimportant and the real 很难从不重要的和真实的中辨别出重要的
from the fake. In this scenario, adversary spoofing, deception and data 从假的。 在这种情况下,对手欺骗,欺骗和数据
manipulation and corruption will create a common operational picture 操纵和腐败将创造一个共同的操作图景
that is part fact, part fiction. This uncertainty in situational awareness will 这部分是事实,部分是虚构。 这种态势感知的不确定性将
feed decision cycles that will be compressed by pervasive data and near 饲料决策周期,将被普遍的数据和附近压缩
instantaneous communications. 瞬时通信。
With these systemic changes, the character of warfare will clearly 随着这些系统性的变化,战争的性质将明显
change, and these changes could significantly influence the Clausewitzian 变化,而这些变化可能会显着影响Clausewitzian
elements that frame our understanding of war’s nature. However, 构成我们对战争本质理解的要素。 然而,
10
autonomy will change the nature of war in several ways, as suggested: 自治将以几种方式改变战争的性质,如所建议的:10
(a) It could weaken the role of political direction by forcing response (a)它可能通过强迫反应来削弱政治方向的作用
delegation to lower echelons for faster forms of attack. (b) Autonomy 授权到较低的梯队以获得更快的攻击形式。 (b)自治
can lessen the ability of governments to gain the support and legitimacy 可以减少政府获得支持和合法性的能力
of their populations, while making it easier for foreign governments to 他们的人口,同时使外国政府更容易
manipulate their adversary’s populations. (c) Deep learning forms of AI 操纵对手的人口。 (c)人工智能的深入学习形式
will augment the intuition and judgment of experienced commanders. (d) 将增强有经验的指挥官的直觉和判断力。 (d)
Automated technologies could reduce popular support for professional 自动化技术可以减少对专业人员的普遍支持
military institutions, which, paradoxically, could free governments to 自相矛盾的是,军事机构可以让政府自由地
employ force more readily since the political consequences are reduced. (e) 更容易使用武力,因为政治后果减少了。 (e)
As with the earlier ages, friction and uncertainty will endure. (f) The age 与早期时代一样,摩擦和不确定性将持续下去。 (f)年龄
of autonomy can introduce new forms of friction while reducing human 自主性可以引入新的摩擦形式,同时减少人类
factors in tactical contexts. To which, the most significant elements of 战术背景下的因素。 其中,最重要的元素
“war, violence, human factors, and chance, will certainly remain”, and “so "战争,暴力,人为因素和机会,肯定会继续存在",并且"所以
11
too, will fog and friction”. War’s essence as politically directed violence 太,会起雾和摩擦"。 11战争的本质是政治导向的暴力
will remain its most enduring aspect, even if more machines are involved 将保持其最持久的方面,即使涉及更多的机器
at every level. As technology advances at a rapid pace, the nature and 在各个层面。 随着技术的快速发展,自然和
character of war will be changed. In the upcoming military revolution of 战争的性质将被改变。 在即将到来的军事革命中
autonomy, we will have to consider new sources of combat power and 自治,我们将不得不考虑新的战斗力来源和
12
assess how they impact each level of war. 评估它们如何影响每个级别的战争。 12
Given this changing nature of warfare, the US Department of 鉴于这种不断变化的战争性质,美国国防部
Defence (DoD) has developed a multi-decade strategy for applying a 国防(DoD)制定了一项为期十年的战略,用于应用
suite of advanced technologies to nearly every facet of its operations. 一套先进的技术,几乎每一个方面的运作.
In the first phase, the DoD will create a more intelligent force, using 在第一阶段,国防部将创建一个更智能的力量,使用
AI to enhance platforms, munitions and decision processes. As these 人工智能加强平台、弹药和决策过程。 作为这些
technologies mature, the US aims to create a more autonomous force, 技术成熟,美国旨在创造一支更自主的力量,
pairing AI enabled systems with human military personnel to accentuate 将AI启用的系统与人类军事人员配对以突出显示
the strengths of each, enabling faster decisions and better combat 每个优势,使更快的决策和更好的战斗
outcomes. In the more distant future, “swarms” of advanced cognitive 结果。 在更遥远的未来,"成群"的高级认知
robots may redefine combat operations in the battle space, as explained 机器人可能会重新定义战斗空间中的作战行动,正如所解释的那样
in Fig 1 below. 在下面的图1中。
Fig 1: Assessment of the Impact of AI and Robotics on 图1:评估人工智能和机器人技术对
Future US Military Operations 未来的美国军事行动
What makes AI significant is the fact that it can operate in several 人工智能之所以重要,是因为它可以在几个方面运行
dimensions, such as: 尺寸,例如:
It can be used to direct physical objects, such as robotic systems, to 它可用于将物理对象(如机器人系统)引导到
act without human supervision. Whether in tanks, planes or ships, 在没有人监督的情况下采取行动。 无论是在坦克,飞机或船舶,
AI can help reduce the need to use humans, even remotely, or as 人工智能可以帮助减少使用人类的需要,甚至可以远程使用
part of human machine teams. Swarm techniques generally involve 人机团队的一部分。 群体技术一般涉及
the creation of supervised algorithms that direct platforms such as 创建指导平台的监督算法,例如
drones. drones.
AI can assist in processing and interpreting information. Image 人工智能可以协助处理和解释信息. 图像
recognition algorithms can be used for tagging. Project Maven is a 识别算法可用于标记。 项目Maven是一个
US military programme that seeks to develop algorithms to automate 寻求开发算法以实现自动化的美国军事计划
the process of analysing video feeds captured by drones. 分析无人机捕获的视频源的过程。
Intelligence, Surveillance, and Reconnaissance: Command, Intelligence, Surveillance, and Reconnaissance: Command,
Control, Communications, Computers, Intelligence, Surveillance and 控制、通讯、电脑、情报、监察及
Reconnaissance (C4ISR) are reaching new heights of efficiency that 侦察(C4ISR)的效率达到了新的高度。
enable data collection and processing at an unprecedented scale and 以前所未有的规模实现数据收集和处理,
speed. AI is useful in intelligence due to the large data sets available for 速度。 人工智能在智能方面很有用,因为它有大量的数据集。
analysis and helps automate the work of human analysts who currently 分析并帮助自动化目前人类分析员的工作
spend hours sifting through videos for actionable information. 花几个小时筛选视频以获取可操作的信息。
Military Drones for Surveillance: Military units across the world 用于监视的军用无人机:世界各地的军事单位
are employing drones to: (a) channel remote communication, both 使用无人机来:(a)通道远程通信,两者都是
video and audio, to ground troops and to military bases; (b) track 视频和音频,地面部队和军事基地;(b)轨道
enemy movement and conduct reconnaissance in unknown areas 敌人的移动和在未知区域进行侦察
of a war zone; (c) assist with mitigation procedures after a war by (c)在战争结束后,通过
searching for lost or injured soldiers and giving recovery insights for 寻找失散或受伤的士兵,并为
a terrain and; (d) aid with operations like peace-keeping and border (d)协助维持和平和边界等行动
surveillance. While drones help in guarding aerial zones, robots can 监视。 虽然无人机有助于保护空中区域,但机器人可以
be deployed on land to assist soldiers in ground operations. Robot 部署在陆地上,协助士兵进行地面行动。 机器人
fleets function like soldier units and carry out collaborated armed 舰队的功能就像士兵单位一样,进行协同武装
activities using multiple techniques. They are self-reliant, adaptable, 使用多种技术的活动。 他们是自力更生的,适应性强的,
and have their fault tolerant systems, all of which contribute to their 并有他们的容错系统,所有这些都有助于他们的
ability to make and execute decisions swiftly and competently. 能够迅速和胜任地做出和执行决策.
AI-Assisted Decision-Making: AI assisted decision-making could 人工智能辅助决策:人工智能辅助决策可以
help leaders at all levels rapidly design, plan and evaluate operations. 帮助各级领导快速设计、规划和评估运营。
Through a continuous deliberative planning process, the AI could 通过持续的审议规划过程,人工智能可以
update and evaluate the plans against the operational environment. The 根据操作环境更新和评估计划。 该
AI would continuously monitor the environment and warn planners 人工智能将持续监测环境并警告规划者
when assumptions are no longer valid or if there is an opportunity to 当假设不再有效或有机会
improve the plans. The true power of AI will be in the teaming of the 改进计划。 人工智能的真正力量将在于
human mind with the AI mind. This type of man-machine teaming 人类的头脑与人工智能的头脑。 这种人机组合
will combine human strengths of goal setting, creativity, and ethical 将结合人类在目标设定、创造力和道德方面的优势
thinking with AI strengthening the rational thought through self- 人工智能思维通过自我强化理性思维-
taught experience, intuition and deep forecasting. 传授经验,直觉和深刻的预测。
very much for the present.20 Countries are now using armed drones 现在非常重要。 20个国家正在使用武装无人机
in combat—Nigeria, Pakistan, and Turkey have recently done so— 在战斗中-尼日利亚,巴基斯坦和土耳其最近这样做了—
and the chances for sophisticated drone technology getting into the 以及先进的无人机技术进入
21
wrong hands increases exponentially. If a terrorist group can steal or 错手成倍增加。 21如果恐怖组织可以偷窃或
purchase a drone from a rogue state or corrupt military or intelligence 从流氓国家或腐败的军事或情报部门购买无人机
officials, then they could rely on the myriad online videos posted that 官员,那么他们可以依靠发布的无数在线视频
essentially demonstrate how these unmanned systems may be used to 从根本上展示了这些无人系统如何被用来
22
conduct an attack. 进行攻击。 22
Swarms: Specifically, with the armed forces, the advantage lies in 成群:具体来说,与武装部队,优势在于
adopting a ‘swarm mindset’. To argue, such a change would largely 采用"群体心态"。 可以说,这样的变化在很大程度上会
be seen in the movement away from the single, exquisite weapons 在远离单一精致武器的运动中可以看到
platforms to those that are small, cheap, unmanned, expendable and 面向小型、廉价、无人、消耗性和
fast. There are many operational advantages of swarms in terms of 快。 在以下方面,群的操作优势很多。
autonomy, quantity and speed. Unmanned systems can take greater 自主性,数量和速度。 无人系统可以采取更大的
risks by reducing survivability while maintaining lethality and 风险通过降低生存能力,同时保持杀伤力和
increasing deployability. increasing deployability.
Conclusion Conclusion
Big-data analysis and machine-learning algorithms are already available 大数据分析和机器学习算法已经可用
and vastly expand information processing capabilities. AI is already a 并大幅扩展信息处理能力。 AI已经是一个
military reality. Automated decision-making will play an increased role 军事现实。 自动化决策将发挥更大的作用
at every level of the command and control process, from swarming 在命令和控制过程的各个层面,从蜂拥而至
miniature UAVs to the national command authority. Genuine AI in the 微型无人机到国家指挥当局。 真正的AI在
scientific sense may still be years away, but it is not too early to begin 科学意识可能还有几年的时间,但现在开始还不算太早
establishing normative limits for LAWS through IHL and military rules 通过国际人道法和军事规则为法律制定规范性限制
27
of engagement, in anticipation of this eventuality. 在预期这种可能性的情况下,参与。 27
Notes Notes
1. Manju Bansal, “AI is the New Black,” MIT Technology Review, February 27, 2017, at 1. Manju Bansal,"AI是新的黑色,"麻省理工学院技术评论,2017年2月27日,在https://
https://www.technologyreview.com/s/603748/ai-is-the-new-black/. www.technologyreview.com/s/603748/ai-is-the-new-black/。
2. Department of Industrial Policy and Promotion, “Report of the Artificial Intelligence 2. 产业政策及推广部,"人工智能专责小组报告",2017年8月,于http://dipp.nic.in/whats
Task Force”, August 2017, at http://dipp.nic.in/whats-new/report-task-force-artificial- -new/report-task-force-artificial 情报。
intelligence.
3. Ibid., p. 6. 3. Ibid., p. 6.
4. Michael E. Howard and Peter Paret, eds., On War: Carl von Clausewitz (Princeton, NJ: 4. Michael E.Howard和Peter Paret,eds。,On War:Carl von Clausewitz(Princeton,NJ:
Princeton University Press, 1989). 普林斯顿大学出版社,1989年)。
5. F. G. Hoffman, “Exploring War’s Character and Nature: Will War’s Nature Change in the 5. F.G.Hoffman,"探索战争的性质和性质:第七次军事革命中战争的性质会改变吗",Para
Seventh Military Revolution”, Parameters, Vol. 47, No. 4, 2017, p. 23, at https://ssi. meters,Vol。 47,No.4,2017,p.23,在https://ssi... armywarcollege.edu/pubs/parameters
armywarcollege.edu/pubs/parameters/issues/Winter_2017-18/5_Hoffman.pdf /issues/Winter_2017-18/5_Hoffman.pdf
6. Michael C. Horowitz, “Artificial Intelligence, International Competition and the Balance 6. Michael C.Horowitz,"人工智能,国际竞争和权力平衡",德克萨斯国家安全评论,Vol
of Power”, Texas National Security Review, Vol. 1, No. 3, May 2018, pp. 46-47, at 。 1,No.3,May2018,pp.46-47,at file:///C:/Users/JASH/Downloads/TNSR-Vol-1-Iss-3_h
file:///C:/Users/JASH/Downloads/TNSR-Vol-1-Iss-3_Horowitz.pdf. orowitz.pdf。
7. Ibid., p. 47. 7. Ibid., p. 47.
8. Ibid. 8. Ibid.
9. Robert O. Work and Shawn Brimley, “20YY: Preparing for War in the Robotic Age”, 9. Robert O.Work和Shawn Brimley,"20yy:为机器人时代的战争做准备",新美国安全中
Centre for a New American Security, January 2014, at https://fortunascorner.com/wp- 心,2014年1月,https://fortunascorner.com/wp 内容/上传/2014/05/cnas_20yy_workbriml
content/uploads/2014/05/cnas_20yy_workbrimley.pdf. ey.pdf。 10. Kareem Ayoub和Kenneth Payne,"人工智能时代的战略",战略研究杂志,Vol
。 39,No.5-6,2016,第793-819页。 11. 霍夫曼,ñ.5,p.31. 12. 同上。 13. 在Eric Gons等
10. Kareem Ayoub and Kenneth Payne, “Strategy in the Age of Artificial Intelligence”, Journal 人中引用。,编辑。,"人工智能和机器人将如何破坏国防工业",波士顿咨询集团,2018年
of Strategic Studies, Vol. 39, No. 5-6, 2016, pp. 793-819. ,第2页,http://k1.caict.ac.cn/yjts/qqzkgz/zksl/201804/P020180426375927121069。pdf
11. Hoffman, n. 5, p. 31. 。 14. 同上。,第3页。 15. 见"未来战争的作战环境和不断变化的性质",网址为http://ww
w.arcic.army.mil/App_Documents/The-Operational-Environment-and-the 改变未来战争的
12. Ibid.
特征。pdf。 16. Sean Kimmons,"凭借多领域概念,陆军旨在'优势之窗'",美国陆军,2016
13. Cited in Eric Gons et al., eds., “How AI and Robotics will Disrupt the Defense Industry”, 年11月14日,在https://www.army.mil/article/178137/with_multi_ domain_concept_army_
The Boston Consulting Group, 2018, p. 2, at http://k1.caict.ac.cn/yjts/qqzkgz/ aims_for_windows_of_superiority。 17. IC面临着跨机构获取、管理、关联、融合和分析越
zksl/201804/P020180426375927121069.pdf. 来越多的数据的挑战。 IC中的数据以太多不同的格式生成,在太多的断开或无法访问的系统
中生成,没有标准化的结构,也没有对本体的总体一致。 这种情况可能造成收款浪费、缺乏
14. Ibid., p. 3.
及时性、遗漏指示和警告以及对决策缺乏相关性。 18. Cortney Weinbaum和John N.T.Shanah
15. See “The Operational Environment and Changing Character of Future Warfare”, at an,"数据驱动时代的情报",联合部队季刊,Vol。 90,2018,在http://ndupress.ndu.edu/Port
http://www.arcic.army.mil/App_Documents/The-Operational-Environment-and-the- als/68/Documents/
Changing-Character-of-Future-Warfare.pdf.
16. Sean Kimmons, “With Multi-Domain Concept, Army Aims for ‘Windows of Superiority’”,
U.S. Army, November 14, 2016, at https://www.army.mil/article/178137/with_multi_
domain_concept_army_aims_for_windows_of_superiority.
17. The IC is challenged to acquire, manage, correlate, fuse, and analyse ever increasing
amounts of data across agencies. Data in the IC are generated in too many diverse formats,
in too many disconnected or inaccessible systems, without standardised structures and
without overarching agreed upon ontology. This situation risks wasted collections, lack of
timeliness, missed indications and warnings and lack of relevance for decision-making.
18. Cortney Weinbaum and John N.T. Shanahan, “Intelligence in a Data-Driven Age”, Joint
Force Quarterly, Vol. 90, 2018, at http://ndupress.ndu.edu/Portals/68/Documents/