See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/340503792

Artificial Intelligence Application in the Military: The Case of United States

and China

Research · June 2019

CITATIONS READS

6 5,620

1 author:

Gloria Shkurti
SETA Foundation
10 PUBLICATIONS 22 CITATIONS

SEE PROFILE

All content following this page was uploaded by Gloria Shkurti on 08 April 2020.

The user has requested enhancement of the downloaded file.

 ARTIFICIAL INTELLIGENCE
APPLICATION IN THE MILITARY
THE CASE OF UNITED STATES
AND CHINA
GLORIA SHKURTI ÖZDEMIR

ANALYSIS JUNE 2019 NO.51

 ARTIFICIAL INTELLIGENCE
APPLICATION IN THE MILITARY
THE CASE OF
UNITED STATES
AND CHINA

GLORIA SHKURTI ÖZDEMIR

COPYRIGHT © 2019 by SETA
All rights reserved.
No part of this publication may be reprinted
or reproduced or utilized in any form or by any
electronic, mechanical or other means, without
permission in writing from the publishers.

ISBN: X

Printed in Turkey, İstanbul

by Turkuvaz Haberleşme ve Yayıncılık A.Ş., 2019

SETA | SİYASET, EKONOMİ VE TOPLUM ARAŞTIRMALARI VAKFI

Nenehatun Cd. No: 66 GOP Çankaya 06700 Ankara TÜRKİYE
Tel: +90 312 551 21 00 | Faks: +90 312 551 21 90
www.setav.org | [email protected] | @setavakfi

SETA | İstanbul
Defterdar Mh. Savaklar Cd. Ayvansaray Kavşağı No: 41-43
Eyüpsultan İstanbul TÜRKİYE
Tel: +90 212 395 11 00 | Faks: +90 212 395 11 11

SETA | Washington D.C.

1025 Connecticut Avenue, N.W., Suite 1106
Washington D.C., 20036 USA
Tel: 202-223-9885 | Faks: 202-223-6099
www.setadc.org | [email protected] | @setadc

SETA | Kahire
21 Fahmi Street Bab al Luq Abdeen Flat No: 19 Cairo EGYPT
Tel: 00202 279 56866 | 00202 279 56985 | @setakahire

SETA | Berlin
Französische Straße 12, 10117 Berlin GERMANY
Tel: +49 30 20188466
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

TABLE OF CONTENTS

ABSTRACT7

INTRODUCTION8

UNDERSTANDING ARTIFICIAL INTELLIGENCE 9

WHY STATES INVEST IN ARTIFICIAL INTELLIGENCE 12

THE UNITED STATES AND ARTIFICIAL INTELLIGENCE 14

CHINA AND ARTIFICIAL INTELLIGENCE 18

CONCLUSION22

s e ta v. o rg 5
ANALYSIS

ABOUT THE AUTHOR

Gloria Shkurti Özdemir

Gloria Shkurti completed her BA in Political Science and International Relations
in Albania. She finished her Master studies at Sakarya University with her thesis
entitled “A Lethal Weapon that Became the Cure-all for Terrorism: Discursive
Construction of the U.S. Dronified Warfare.” Currently, she is a Ph.D. candidate
at Ankara Yıldırım Beyazıt University. Her main research interests include drone
warfare, artificial intelligence, U.S. foreign policy, and security studies. Currently,
she is working as the Assistant to Editor-in-Chief of Insight Turkey, a journal pub-
lished by SETA Foundation.

6 s e ta v. o rg
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

ABSTRACT

This analysis provides a short introduction on what

Artificial Intelligence (AI) is, how it has evolved until
today, and how it will change the nature of warfare.

Considered as the 4th Industrial Revolution, Artificial Intelligence (AI) has

become a reality in today’s world, especially in the military. Experts and aca-
demicians have emphasized the importance of AI for a long time. Further-
more, world leaders, including Obama, Trump, Xi, and Putin, have all made
important statements that bring to the fore the significance of AI which can
be summarized with what Putin stated on September 2017: whoever be-
comes the leader in AI, will rule the world.
This analysis provides a short introduction on what AI is, how it has
evolved until today and how it will change the nature of warfare. It then as-
sesses why states invest in AI to later turn to the case of the U.S. and China.
For both states, the main official documents and statements are analyzed, the
bureaucratic structures that work on AI are presented and finally examples
of how the U.S. and China are applying AI in the military are provided. The
conclusion briefly comments on how the strategies of China and the U.S.
differ followed by some recommendation on what states like Turkey should
do in the near future.

s e ta v. o rg 7
ANALYSIS

For this reason, many state leaders, technol-

ogy experts, and academicians have considered
AI to be revolutionary; yet they differ in terms of
their perspectives on AI. Specifically, some con-
sider AI as a positive development as it would
help in reducing human causalities -considering
that the human force would be replaced by the
machines which can be deployed in every kind of
mission, including the high risk ones- and at the
same time it would provide strategic and tactical
advantages. However, others caution on the fact
that if not controlled and used properly, AI may
result in another world war.
Even though AI is still at its juvenile stage, it
INTRODUCTION is undeniable that it has the capacity to alter the
Artificial Intelligence (AI) has advanced very landscape of the security sector and as a result, it
quickly in recent years and this has resulted in a will change the current economic and military
wide range of applications, both civil and military. balances in the international system. Recogniz-
It is a fact that the military is driven by the need ing the potentials of AI, more than 20 countries
for change as it is persistently on the pursuit of have announced their national AI strategies,
better, faster and stronger weapons or technolo- and more states and non-state organization are
gies, and this is exactly what AI provides. As is taking decisive steps in AI research and devel-
correctly stated by Svenmarck et al., in the field opment (R&D). Yet, the U.S. and China are
of military, AI has the potential to impact “all do- considered to be the leading states in the field;
mains (i.e. land, sea, air, space and information) the U.S. aiming to remain the hegemon in the
and all levels of warfare (i.e. political, strategic, op- battlefield, while China is aiming to leapfrog
erational and tactical).”1 Currently, the use of AI the U.S. and become the leader in AI by 2030.
in the military is seen mainly in terms of: Needless to say, both these states have generat-
• Autonomous Weapons and Weapons Tar- ed independent national strategies that comply
geting with these aims, but most importantly, the U.S.
• Surveillance national strategy takes as its main reference point
the Chinese national strategy, and vice versa.
• Cybersecurity
This analysis provides a short introduction
• Homeland Security
on what AI is, how it has evolved until today, and
• Logistics how it will change the nature of warfare. It then
• Autonomous Vehicles2 assesses why states invest in AI to later turn to the
case of the U.S. and China. For both states, the
1 Peter Svenmarck et al., “Possibilities and Challenges for Artificial In-
telligence in Military Applications,” (n.d.), 1, https://www.sto.nato.int/
main official documents and statements are ana-
publications/.../STO-MP-IST-160/MP-IST-160-S1-5.pdf. lyzed, the bureaucratic structures that work on AI
2 Marcus Roth, “Artificial Intelligence in the Military – An Overview of
Capabilities,” Emerj, February 22, 2019, https://emerj.com/ai-sector-over-
are presented and examples are provided of how
views/artificial-intelligence-in-the-military-an-overview-of-capabilities/. the U.S. and China are applying AI in the military.

8 s e ta v. o rg
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

In the conclusion section, there is a brief comment ability of “big data” sources, (ii) improvements to
on how the strategies of China and the U.S. dif- machine learning approaches, and (iii) increases
fer, followed by some recommendations on what in computer processing power.8 Currently, we
states like Turkey should do in the near future. can speak about two types of AI: Narrow AI
(NAI) and General AI (GAI). While NAI refers
to functions like playing games or image recog-
UNDERSTANDING nition, GAI is used to denote systems capable
ARTIFICIAL INTELLIGENCE of human-level intelligence which are adequate
Considered as the 4th Industrial Revolution,3 AI to conduct a series of different tasks. Currently,
has become a reality in today’s world and many all the known AI developments fall within NAI
experts argue that AI itself should not be consid- while in terms of GAI, experts are generally skep-
ered as a specific weapon but it should be seen tical and it is believed that decades are needed
as “an enabler, a general-purpose technology until it reaches its full capability.9 Yet in order
with a multitude of applications.”4 So while “AI to understand better, this classification is not
could potentially enable a number of military in- important. An important issue related to AI, re-
novations, it is not a military innovation itself.”5 gardless of whether it is narrow or general, is that
Kevin Kelly, a technology expert, compared AI of the autonomy. The next subsection provides
with electricity stating that “just as electricity a brief outline in terms of the spectrum of au-
brings objects all around us to life with power, so tonomy which results from the use of AI.
too will AI bring them to life with intelligence.”6
However, technology experts are not the only Types of Autonomy:
ones that emphasize the importance of AI. Human-Machine Relationship
World leaders, including Obama, Trump, Xi and When it comes to the application of AI, especially
Putin, have all made important statements that in the military, the human-machine relationship
bring to the fore the significance of AI which can is one of the main topics discussed. Currently,
be summarized with what Putin stated on Sep- there are three types of relationships that we can
tember 2017: whoever becomes the leader in AI, debate about: (i) human in the loop, (ii) human
will rule the world.7 on the loop, and (iii) human out of the loop.
AI has been developing since 1956, but in- When the human is in the loop it means that
terest in this field started to increase circa 2010 the machine is in the control of the environment;
due to three enabling developments: (i) the avail- however, it is the human that takes the final deci-
sion. This is called as a semi-autonomous system.
3 Paul Scharre, Army of None: Autonomous Weapons and the Future of War, In the second case, that of human on the loop, the
Kindle Edition (New York, London: W. W. Norton & Company, 2018),
16; Klaus Schwab, The Fourth Industrial Revolution (Davos: World Eco- machine can act and decide on its own. Neverthe-
nomic Forum, 2016), 12.
4 Michael C. Horowitz, “Artificial Intelligence, International Competi- 8 Kelley M. Sayler, “Artificial Intelligence and National Security” (Con-
tion, and the Balance of Power,” Texas National Security Review 1, no. 3 gressional Research Service, January 30, 2019), 2.
(May 2018): 39. 9 One of the latest surveys published in 2018, based on the expert’s opin-
5 Horowitz, 43. ion argued that there was “a 50% chance of AI outperforming humans
in all tasks in 45 years.” Some others doubted GAI is possible and many
6 Scharre, Army of None: Autonomous Weapons and the Future of War, 16. who did support its chances put the date hundreds of years in the future.
7 “Whoever Leads in AI Will Rule the World’: Putin to Russian Children See, Katja Grace et al., ‘Viewpoint:
on Knowledge Day,” RT International, accessed April 21, 2019, https:// When Will AI Exceed Human Performance? Evidence from AI Experts’,
www.rt.com/news/401731-ai-rule-world-putin/. Journal of Artificial Intelligence Research, 62 (2018), 729-754.

s e ta v. o rg 9
ANALYSIS

less, the human has the possibility to observe the this is the American Project Maven11 that is being
behavior of the machine and intervene if neces- programed to process and interpret the informa-
sary. In this case, we have a supervised autono- tion received from the videos captured by drones.
mous system. The last case is that of a fully au- (ii) By using AI, drones, planes, ships, tanks, etc.
tonomous system. Here, while the machine acts will not need a human to control or direct their
and decides by itself, the human does not have actions. An example of this is the Israeli Harpy
any control on the machine, and as a result re- drone that has passed to a close to full autonomy
mains out of the loop. Currently, in the military (it still needs to be launched by ground troops).
application we can speak about the application of Applying a lesser level of autonomy, other aircrafts
the first or second case, i.e. drones and precision- such as Air Force Global Hawk and Army Gray
guided missiles, while the AI application has not Eagle drones can be mentioned. For both these
yet reached the third level of full autonomy. drones, it is necessary that the operator directs
the aircraft where to go and then the aircraft flies
AI Military Application itself. As a result, it is expected that autonomous
The rapid development of artificial intelligence has systems will replace humans in tasks that are
been followed with positive impacts in different considered as “dull, dangerous, or dirty.”12 Such
fields, including medicine or transportation; how- tasks may include long-duration intelligence col-
ever, when it comes to the military application a lot lection and analysis, clean up environments con-
of discussions are raised. It is widely accepted now, taminated by chemical weapons, or sweep routes
that in the future, AI has the capacity to transform for improvised explosive devices. (iii) Lastly, as
military activities such as logistics, intelligence, algorithms develop further, AI may be used for
surveillance and especially weapons design.10 For command and control, including battle manage-
the time being, the uncertainties that encircle AI ment, by analyzing large sets of data and making
and its use have raised a lot of debates among mili- forecasts to direct human action.
tary strategists. These debates are generally related
to the impact of AI on the way in which warfare AI’s Impact on the Character of
is conducted and how much autonomy should be the Warfare
given to the weapons that are going to use AI. In As mentioned earlier, much discussion is con-
particular, the main discussions are with regard to centrated on AI’s application in the military. On
the lethal autonomous weapons systems (LAWS) the one hand stand those scholars that approach
or “killer robots” as they are dubbed. These sys- the application of AI skeptically, as they bring
tems fall within the third case mentioned earlier, to the fore the fact that humans can lose control
where the human is out of the loop. over the machines which would wreak havoc.13
Currently, it is argued that in terms of mili-
11 Zachary Fryer-Biggs, “In Project Maven’s Wake, the Pentagon Seeks
tary application, AI can be used (but not limited
AI Tech Talent,” Wired, December 21, 2018, https://www.wired.com/
to) as follows: (i) Based on the image-recognition story/inside-the-pentagons-plan-to-win-over-silicon-valleys-ai-experts/.

algorithms, AI can be used to assist in process- 12 Mick Ryan, “Integrating Humans and Machines,” The Strategy Bridge,
January 2, 2018, https://thestrategybridge.org/the-bridge/2018/1/2/
ing and interpreting information. An example for integrating-humans-and-machines.
13 Paul Scharre, “Autonomous Weapons and Operational Risk” (Center
for a New American Security, 2016), 35, https://s3.amazonaws.com/files.
10 Kenneth Payne, “Artificial Intelligence: A Revolution in Strategic Af- cnas.org/documents/CNAS_Autonomous-weapons-operational-risk.
fairs?,” Survival: Global Politics and Strategy 60, no. 5 (2018): 9. pdf?mtime=20160906080515.

10 s e ta v. o rg
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

Furthermore, there is also the possibility that resist the application of AI in the military.19 The
not only states, but also the non-state actors,14 experts that fall within the second group, those
including here the terrorist organizations, will who believe in the evolutionary impact of AI, ac-
be able to use the AI systems further deepening cept the fact that AI will play a crucial role in
the asymmetry of warfare. On the other hand, warfare -even though GAI is unlikely to develop
AI machines are seen as a positive development in the near future- yet they call for the impor-
as they would replace the human soldiers in dan- tance of the human presence as this would keep
gerous combat and as a result the human loses the AI from taking control over the warfare.20 The
will be less.15 Moreover, these machines may last group of experts is the one that supports the
prove more beneficial in long-duration opera- idea that AI will have a revolutionary impact on
tions where humans cannot endure.16 warfare. Proponents of this position contend that
However, despite the pros-and-cons debate AI’s application in the military has the capability
on AI warfare applications, experts have made to change the nature and principles of warfare.
predictions regarding the impact scale of AI on Accordingly, they speak about a “transition from
the future of the character of warfare. In this re- the industrial era of warfare into the information
spect, there are three possible positions: (i) mini- era, in which gathering, exploiting, and dissemi-
mal impact, (ii) evolutionary impact, and (iii) nating information will be the most consequen-
revolutionary impact.17 tial aspect of combat operations.”21 At this point,
Those experts who argue on the minimal AI’s capability to process a large amount of data
impact are mainly concerned with the technical in a very short time will be a crucial wartime ad-
aspect of AI and the military organizations’ will- vantage, allowing quicker and better decisions.22
ingness to use this technology. Accordingly, they
contend that the problems that could result from Many states have already
the use of AI will make it incompatible for mili- made public their AI
tary applications.18 Regarding military organiza- national strategies.
tions, some bureaucracies are not open to radi-
cal changes such as AI and as a result they would Despite these discussions, no one doubts
the fact that the character of warfare will be
changed. First of all, warfare will be pushed to

14 Thomas X. Hammes, “Technology Converges; Non-State Actors 19 Gautam Mukunda, “We Cannot Go On: Disruptive Innovation and
Benefit,” Governance In An Emerging New World, no. 319 (February the First World War Royal Navy,” Security Studies 19, no. 1 (2010): 136;
25, 2019), https://www.hoover.org/research/technology-converges-non- Barry R. Posen, The Sources of Military Doctrine: France, Britain, and
state-actors-benefit. Germany Between the World Wars (Cornell: Cornell University Press,
15 Robert O. Work and Shawn Brimley, “20YY Preparing for War in 1986).
the Robotic Age” (Washington, DC: Center for a New American Secu- 20 Work and Brimley, “20YY Preparing for War in the Robotic Age,”
rity, 2014), 7, https://fortunascorner.com/wp-content/uploads/2014/05/ 25; Kareem Ayoub and Kenneth Payne, “Strategy in the Age of Artificial
cnas_20yy_workbrimley.pdf. Intelligence,” The Journal of Strategic Studies 39, no. 5 (2015): 816.
16 Office of Technical Intelligence, “Technical Assessment: Autonomy” 21 Sayler, “Artificial Intelligence and National Security,” 35.
(Washington, DC: Department of Defense, 2015), 6, https://apps.dtic.
mil/dtic/tr/fulltext/u2/a616999.pdf. 22 John R. Allen and Amir Husain, “On Hyperwar,” Proceedings
143/7/1,373 (2017): 30; “Getting to Grips with Military Robotics - War
17 Sayler, “Artificial Intelligence and National Security,” 33–36. at Hyperspeed,” The Economist, January 25, 2018, https://www.econo-
18 Paul Scharre and SSQ, “Highlighting Artificial Intelligence: An In- mist.com/special-report/2018/01/25/getting-to-grips-with-military-
terview with Paul Scharre Director, Technology and National Security robotics; Williamson Murray and MacGregor Knox, “The Future Behind
Program Center for a New American Security Conducted 26 September Us,” in The Dynamics of Military Revolution, 1300-2050 (Cambridge:
2017,” Strategic Studies Quarterly 11, no. 4 (2017): 17. Cambridge University Press, 2001), 178.

s e ta v. o rg 11
ANALYSIS

the limits in terms of time scale. AI’s capacity to mission, in 2016 there were investments in AI
react on machine speed will accelerate the pace of approximately €10 billion in Asia and €18
of combat.23 Whether this development will be billion in the U.S..26 This amount may be even
beneficial or not is highly contested among ana- larger considering that not all states declare their
lysts.24 Second, the current military structure and funding in AI. Yet, according to Scharre, in mili-
organizations are going to change as new con- tary robotics alone, the global spending in 2018
cepts of operations evolve (ex: swarm drones). is expected to be $7.5 billion per year.27
Third, AI may provide the opportunity to cope AI is now considered as the leverage that
with a huge amount of data available for analy- would give advantages to those states which use
sis. By addressing this data, AI systems will be it, not only in the civil aspects of life but also
able to provide results or solutions that humans (and most importantly) in the military. The ad-
may not be able to think about,25 especially when vent of AI and its application would create great
found in combat. This would then serve as an opportunities for the states that use it first, i.e.
advantage over the enemy. Lastly, a new debate the U.S. and China, it will particularly help these
on the quality-quantity issue is going to emerge states to gain economic and military advantages
considering the feasibility of achieving a specific over their competitors.28 Furthermore, it is ex-
AI system once the software is created. pected that it will redefine the current balances
All these discussions have not stopped the of power. However, the repercussions would be
states from pursuing the application of AI in dif- detrimental for the states that fail to adopt this
ferent aspects of life, including the military. In technology in their military. Currently, Turkey is
what has started to seem more like a race, many among the states that have not taken consider-
states have already made public their AI national able steps in the area of AI R&D. Yet, consider-
strategies. In this connection, the next section of ing the current regional dynamics and the fact
this analysis, will deal with the main reason why that states like UAE and Israel have started to
states are willing to invest in AI. focus on AI, it is important that Turkey creates
its own AI national strategy and allocates consid-
erable funding to this area.
WHY STATES INVEST IN The main reason behind the power of AI to
ARTIFICIAL INTELLIGENCE shift the current balances of power relies on the
Currently, as can be seen in Table 1, approxi- fact that the application of AI in the military will
mately 20 states/organizations have announced give the states the upper hand in the battlefield,
or are about to announce their AI strategies. as the machines will be more accurate and faster
Furthermore, according to the European Com- than humans in logistics, battlefield and deci-
sion making. Furthermore, with the help of AI,
23 Greg Allen and Taniel Chan, “Artificial Intelligence and National Se- the military can perform high risk missions for
curity” (Belfer Center for Science and International Affairs, 2017), 24.
24 Scharre, “Autonomous Weapons and Operational Risk”; Peter W.
Singer, Wired for War: The Robotics Revolution and Conflict in the Twenty- 26 Florin Zubașcu, “Commission Says €9.2 Billion Needed to Boost
First Century (New York: Penguin Press, 2009), 128. Supercomputing and AI,” Science|Business, June 6, 2018, https://sci-
encebusiness.net/news/commission-says-eu92-billion-needed-boost-
25 The AI company DeepMind created a game-playing algorithm called supercomputing-and-ai.
AlphaGo. In March 2016 Alpha Go defeated a world- champion Go
player, Lee Sedol, 4-1. Sedol stated that AlphaGo made surprising and 27 Scharre, Army of None: Autonomous Weapons and the Future of War.
innovative moves, and other expert Go players stated that AlphaGo over- 28 Paul Scharre, “Killer Apps: The Real Dangers of AI Arms Race,” For-
turned accumulated wisdom on game play. eign Affairs 98, no. 3 (June 2019): 139.

12 s e ta v. o rg
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

TABLE 1: LIST OF THE STATES WHICH HAVE/ARE ABOUT TO ANNOUNCE THEIR

AI NATIONAL STRATEGY
No. Date State National Strategy Title
1 March 2017 Canada Pan-Canadian AI Strategy
2 March 2017 Japan Artificial Intelligence Technology Strategy
3 May 2017 Singapore AI Singapore
4 July 2017 China A Next Generation Artificial Intelligence Development Plan
5 October 2017 UAE UAE Strategy for Artificial Intelligence
6 December 2017 Finland Finland’s Age of Artificial Intelligence
7 January 2018 Kenya Blockchain & Artificial Intelligence Taskforce
8 January 2018 Taiwan Taiwan AI Action Plan
9 January 2018 Denmark Strategy for Denmark’s Digital Growth
10 March 2018 Italy Artificial Intelligence at the Service of Citizens
11 March 2018 France France’s Strategy for AI
National AI Strategy: Unlocking Tunisia’s capabilities potential (Work-
12 April 2018 Tunisia
shop – NS to be announced)
13 April 2018 EU Commission Communication on Artificial Intelligence for Europe
14 April 2018 UK Industrial Strategy: Artificial Intelligence Sector Deal
15 May 2018 Australia Australian Technology and Science Growth Plan
16 May 2018 South Korea Artificial Intelligence R&D Strategy
17 May 2018 Sweden National Approach for Artificial Intelligence
18 June 2018 India National Strategy for Artificial intelligence: #AIforAll
19 June 2018 Mexico Towards an AI Strategy in Mexico: Harnessing the AI Revolution
20 December 2018 Germany Key points for a Federal Government Strategy on AI
21 February 2019 U.S. Accelerating America’s Leadership in Artificial Intelligence

a long period of time, something that cannot be Strategy, AI is considered to be cost effective.29 The
done by humans. possibility to create or buy cheap robots drives the
Moreover, especially for democracies, AI states to consider the quantity over the quality
will make the legitimation of war easier by the and such a policy, at least in the near future will
governments as it reduces the human participa- be translated into an advantage. Yet, it needs to be
tion in the battlefield and as a result minimizes highlighted that AI cannot remain in the hands of
the human causalities. For autocratic regimes, just a couple of states, such as the case with nuclear
i.e. China, application of AI in the military weapons. Soon enough, a considerable number of
would help the government to concentrate the states or even non-state organizations will be able
power to conduct war in only a small trustable to integrate AI into their military. At this point the
group of people, which would be translated into quantity will not generate any advantage anymore
greater control of the war by the elites. so the states or non-state actors will start to focus
Lastly, one of the main reasons why the on the quality rather than quantity.
states are eager to integrate AI into their militaries
is related to the cost. As mentioned even in the 29 “The U.S. Army Robotic and Autonomous Systems Strategy” (U.S.
U.S. Army’s Robotics and Autonomous Systems Army, March 2017), 3, https://www.tradoc.army.mil/Portals/14/Docu-
ments/RAS_Strategy.pdf.

s e ta v. o rg 13
ANALYSIS

THE UNITED STATES AND AI national strategy. It is necessary to emphasize

the fact that the U.S. was among the later states to
ARTIFICIAL INTELLIGENCE announce a formal AI national strategy while Can-
The United States considers AI as a part of its ada was the first state to do so in 2017.34 External
Third Offset Strategy30 which was launched in constraints can be seen in terms of the challenges
2014 by the then Secretary of Defense Chuck that other states, mainly China and Russia, pose to
Hagel with the aim of reviving America’s military the American leadership in AI. Both of these states
technological advantage.31 Within this regard, the aim at challenging the U.S. military superiority
main focus of the Third Offset Strategy is robot- by using AI to generate weapons superior to those
ics and autonomy where AI plays a crucial role.32 used by the U.S.35 Therefore, this section will brief-
Following this, both the Obama and Trump Ad- ly analyze the main national governmental initia-
ministrations have expressed their intentions for tives by the Obama and Trump Administrations,
the U.S. to be the global leader in AI. On May the organizations responsible for the research and
2018, the Trump Administration stated that in delivery of the AI and the main application of AI
order to achieve this goal “to the greatest degree in military until now.
possible, we will allow scientists and technologists The first declaration regarding AI was
to freely develop their next great inventions right made in 2016 when the Obama Administra-
here in the United States.” 33 Nevertheless, despite tion released a road map through which the
these intentions it can be said that the U.S. faces importance of U.S. leadership in AI was ac-
both domestic and external constraints on its way knowledged. However, this roadmap was main-
to becoming the world leader in AI. ly focused on regulatory policy questions.36
First of all, the domestic constraints result Within the same year, the White House Office
from the fact that AI, unlike nuclear weapons, is of Science and Technology Policy organized dif-
being developed by non-governmental institu- ferent workshops37 and created the Subcommit-
tions, such as Google or Microsoft, not by the gov- tee on Machine Learning and Artificial Intel-
ernment itself. As will be discussed below, these in- ligence (MLAI).38 The main goal was to help in
stitutions are not always willing to cooperate with
34 Jessica Baron, “Will Trump’s New Artificial Intelligence Initiative
the government, which impedes the application
Make The U.S. The World Leader In AI?,” Forbes, February 11, 2019,
of AI in the military. Moreover, the government https://www.forbes.com/sites/jessicabaron/2019/02/11/will-trumps-
new-artificial-intelligence-initiative-make-the-u-s-the-world-leader-in-
itself has been moving very slowly in terms of an ai/.
35 William Carter, Emma Kinnucan, and Josh Elliot, “A National Ma-
30 The First Offset Strategy was launched in 1950s and is related to the chine Intelligence Strategy or the United States” (Washington, DC:
U.S.’ investments to nuclear weapons. The Second Offset Strategy is fo- Center for Strategic and International Studies (CSIS), 2018), 17; Sayler,
cused on precision-guided weapons and was developed in 1970s. “Artificial Intelligence and National Security,” i.
31 Hagel Chuck, “Secretary of Defense Speech: Reagan National Defense 36 Barack Obama, “The Administration’s Report on the Future of Ar-
Forum Keynote,” U.S. Department of Defense, November 15, 2014, tificial Intelligence,” Obama White House, October 12, 2016, https://
https://dod.defense.gov/News/Speeches/Speech-View/Article/606635/ obamawhitehouse.archives.gov/blog/2016/10/12/administrations-re-
reagan-national-defense-forum-keynote/. port-future-artificial-intelligence.
32 Jesse Ellman, Lisa Samp, and Gabriell Coll, “Assessing the Third 37 “Preparing for the Future of Artificial Intelligence,” Obama
Offset Strategy” (Washington, DC: Center for Strategic and Interna- White House, May 3, 2016, https://obamawhitehouse.archives.gov/
tional Studies (CSIS), March 2017), https://csis-prod.s3.amazonaws. blog/2016/05/03/preparing-future-artificial-intelligence.
com/s3fs-public/publication/170302_Ellman_ThirdOffsetStrategySum- 38 Executive Office of the President of the United States, “Charter of
mary_Web.pdf. the Subcommittee on Machine Learning and Artificial Intelligence,
33 David Shepardson, “Trump Administration Will Allow AI to ‘freely Committee of Technology, National Science and Technology Council”
Develop’ in U.S.:...,” Reuters, May 10, 2018, https://www.reuters.com/ (Obama White House, n.d.), https://obamawhitehouse.archives.gov/
article/us-usa-artificialintelligence-idUSKBN1IB30F. sites/default/files/microsites/ostp/NSTC/ai_charter_-_signed_final.pdf.

14 s e ta v. o rg
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

the coordination of the national AI activity and the character of war.”44 In a similar way, in its
as a result of these developments three main re- AI strategy report, the Department of Defense
ports emerged under the Obama Administra- calls for the immediate need for the adoption
tion: (i) Preparing for the Future of Artificial of a national AI strategy because “[other] na-
Intelligence,39 (ii) The National Artificial Intel- tions, particularly China and Russia, are mak-
ligence Research and Development Strategic ing significant investments in AI for military
Plan,40 and (iii) Artificial Intelligence, Auto- purposes, including in applications that raise
mation, and the Economy.41 The first report questions regarding international norms and
recommended possible strategies related to AI human rights. These investments threaten to
regulations, public research and development erode our technological and operational advan-
(R&D), ethics, and security. The National Ar- tages and destabilize the free and open interna-
tificial Intelligence Research and Development tional order.”45
Strategic Plan was focused mainly on the R&D
strategy, while the last report pondered over the Both the Obama and Trump
necessary policies that would intensify the ben- administrations have expressed
efits of AI and alleviate its costs. their intentions for the U.S.
Another important step was taken by the to be the global leader in AI.
Department of the Homeland Security in 2017
with the publication of a narrative analysis on In May 2018, the Deputy Assistant to
AI.42 This report focused on how to better un- the President for Technology Policy, Michael
derstand the perception of benefits and threats Kratsios, during the 2018 White House Sum-
from the adoption of AI. Within the same year mit on Artificial Intelligence for American In-
AI was mentioned for the first time in the Na- dustry recapped the Trump Administration’s
tional Security Strategy (2017). The main focus AI goals as such: (i) prioritizing funding for
here was on the role of AI in information state- AI R&D; (ii) removing barriers to innovation;
craft, weaponization, and surveillance.43 Later (iii) training the future American workforce;
on in 2018, the National Defense Strategy stat- (iv) achieving strategic military advantage to
ed that AI is among the “very technologies that lead in AI; (v) leveraging AI for government
ensure we [the U.S.] will be able to fight and services; and (vi) leading international AI ne-
win the wars of the future…. New commercial gotiations.46 In order to achieve all these goals,
technology will change society and, ultimately, Kratsios during the summit announced the
creation of a new Select Committee on Artifi-
cial Intelligence to “improve the coordination
39 “Preparing for the Future of Artificial Intelligence” (Washington, DC:
Executive Office of the President of the United States, 2016).
40 National Science and Technology Council, “The National Artificial 44 “Summary of the 2018 National Defense Strategy of the United States
Intelligence Research and Development Strategic Plan” (Executive Office of America: Sharpening the American Military’s Competitive Edge”
of the President of the United States, 2016). (Washington, DC: Department of Defense, 2018), 3.
41 “Artificial Intelligence, Automation, and the Economy” (Washington, 45 “Summary of the 2018 Department of Defense Artificial Intellgence
DC: Executive Office of the President of the United States, 2016). Strategy: Harnessing AI to Advance Our Security and Prosperity” (Wash-
42 “Narrative Analysis: Artificial Intelligence” (Washington, DC: De- ington, DC: Department of Defense, 2018), 5.
partment of Homeland Security, 2017). 46 “2018 White House Summit on Artificial Intelligence for American
43 Donald J. Trump, “National Security Strategy of the United States Industry” (The White House Office of Science and Technology Policy,
2017” (The White House, 2017). October 5, 2018).

s e ta v. o rg 15
ANALYSIS

of Federal efforts related to AI and ensure con- Bureaucratic Structure

tinued U.S. leadership in AI.”47 While it is not easy to talk about the funding
Just a few months later, in July 2018, the Ex- structure of AI due to the lack of information,
ecutive Office of the President released a memo on we can to some extent list some governmental
FY2020 Administration Research and Develop- institutions responsible for the AI R&D in the
ment Budget Priorities. The memo considered AI U.S. The main responsible institution is the De-
as one of the highest R&D priorities for FY2020 partment of Defense. In June 2018, the Pentagon
and emphasized the importance of AI leadership formed the Joint Artificial Intelligence Center
(alongside the quantum information sciences and (JAIC). According to the director of JAIC, Lt.
strategic computing).48 Within a similar frame- Gen. Jack Shanahan, JAIC is crucial in transi-
work, this memo was followed by an Executive tioning from research and development to oper-
Order signed in February 2019, in which it was ational-fielded capabilities.52 In the last FY2020
stated that the U.S. is the world leader in AI re- proposal prepared by the Trump Administration,
search and maintaining this leadership is of para- the JAIC is allocated $208 million.53 Another
mount importance for the U.S.49 Lastly it is neces- agency responsible for AI projects within the
sary to highlight the fact that the first time that Department of Defense is the Defense Advanced
AI was considered as a governmental priority was Research Projects Agency (DARPA) which has
in the FY2019 budget request of the Trump Ad- taken the lead in terms of AI R&D. DARPA re-
ministration where it was stated that the “budget’s quested a $3.17bn budget in 2018, and $3.44bn
key areas of focus include artificial intelligence, au- in 2019.54 Furthermore, in 2018 DARPA de-
tonomous systems, and hypersonics.”50 clared a multi-year investment of more than
Despite the considerable number of official $2bn in new and existing programs known as
statements and documents released by both the the “AI Next” campaign.55
Obama and Trump Administrations, they lack a Lastly, the Trump White House has estab-
clear funding and bureaucratic structure. This as- lished the Select Committee on Artificial Intel-
pect of the American stance is highly criticized by ligence under the National Science and Technol-
experts who warn of the Chinese rivalry and the ogy Council on May 10, 2018.56 This committee
challenges it is posing to the American leadership.51 serves as the main advisor to the White House
47 “2018 White House Summit on Artificial Intelligence for American
regarding the AI R&D priorities and at the same
Industry,” 7.
48 Executive Office of the President of the United States, “Memorandum
for the Heads of Executive Departments and Agencies: FY 2020 Admin- 52 Billy Mitchell, “Pentagon Unveils Strategy for Military Adoption of
istration Research and Development Budget Priorities” (Washington, Artificial Intelligence,” FedScoop, February 12, 2019, https://www.fed-
DC: The White House, July 31, 2019), https://www.whitehouse.gov/ scoop.com/artificial-intelligence-pentagon-military-unclassified/.
wp-content/uploads/2018/07/M-18-22.pdf. 53 “Budget of the United States Government, Fiscal Year 2020” (Wash-
49 Executive Office of the President of the United States, “Executive ington, DC: The White House, November 3, 2019), 24, https://www.
Order on Maintaining American Leadership in Artificial Intelligence,” govinfo.gov/content/pkg/BUDGET-2020-BUD/pdf/BUDGET-
The White House, accessed May 5, 2019, https://www.whitehouse.gov/ 2020-BUD.pdf.
presidential-actions/executive-order-maintaining-american-leadership- 54 Sebastian Moss, “Understanding the United States’ National AI Strat-
artificial-intelligence/. egy,” February 12, 2019, https://www.datacenterdynamics.com/analysis/
50 “An American Budget: Fiscal Year 2019” (Washington, DC: Office of understanding-united-states-national-ai-strategy/.
Management and Budget, 2019), 36. 55 “DARPA Announces $2 Billion Campaign to Develop Next Wave of
51 Thomas H. Davenport, “China Is Catching up to the US on Artifi- AI Technologies,” DARPA, July 9, 2019, https://www.darpa.mil/news-
cial Intelligence Research,” The Conversation, February 27, 2019, http:// events/2018-09-07.
theconversation.com/china-is-catching-up-to-the-us-on-artificial-intelli- 56 “2018 White House Summit on Artificial Intelligence for American
gence-research-112119. Industry,” 7.

16 s e ta v. o rg
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

time is responsible for the collaboration of the to each other like swarms in nature.”60 A similar
government with industry and academia regard- experiment was done in November 2016 by the
ing AI research and development. Navy. A swarm of five uninhabited boats pa-
trolled a specific section in Chesapeake Bay and
AI Military Applications in the U.S. stopped an “intruder” vessel. Other experiments
AI is a field of technology which has consequen- with swarms of underwater drones are expected
tial implications for the national security. As to be conducted in the near future.
such, the U.S. and other nations are develop- One of the most well-known and discussed
ing AI applications for military functions. Many cases, in which the U.S. government is involved,
different experiments have been conducted in is the Project Maven which incorporates AI in in-
this field and it can be said that until now no telligence, surveillance, and reconnaissance appli-
other state, including here China or Russia, have cations. Until now Project Maven has been used
reached the level where the U.S. is currently. To to support the war against ISIS in Iraq and Syria.61
demonstrate this, this subsection will detail some
of the main test/experiments undertaken by the It can be said that until now no
U.S., especially within the scope of the Depart- other state, including here China
ment of Defense.
or Russia, have reached the level
In 2013, the X-47B prototype drone of the
U.S. Navy landed autonomously.57 The same
where the U.S. is currently.
drone in 2015 conducted autonomous aerial re- Furthermore, the U.S. military services are
fueling.58 In both of the cases, the only human focusing on the incorporation of the AI into semi-
input was the order to land or refuel while in autonomous and autonomous vehicles, including
flight respectively and the action was carried out fighter aircraft, drones, ground vehicles, and naval
by the software. vessels. The Loyal Wingman program is an exam-
As mentioned previously the military orga- ple of this. On the basis of this program, an older-
nization is going to change as AI is applied in the generation uninhabited fighter jet (F-16 or B-1) or
military and new concepts of operation evolve. a low-cost uninhabited aircraft (XQ-58A Valkyrie)
An example for this was the swarm drones. In were paired with an inhabited fighter jet (F-35 or
2016, the U.S. demonstrated 103 drones flying F-22).62 In this case, the uninhabited aircraft by
together autonomously.59 Pentagon described acting autonomously and without preprograming
this as “a collective organism, sharing one dis- has the duty to protect the inhabited jet. Similar
tributed brain for decision-making and adapting
60 Thomas Gibbons-Neff, “Watch the Pentagon’s New Hive-Mind-
Controlled Drone Swarm in Action,” Washington Post, Octo-
57 Raya Jalabi and Spencer Ackerman, “US Navy Makes History by ber 1, 2017, https://www.washingtonpost.com/news/checkpoint/
Landing Unmanned Drone on Aircraft Carrier,” The Guardian, Octo- wp/2017/01/10/watch-the-pentagons-new-hive-mind-controlled-
ber 7, 2013, https://www.theguardian.com/world/2013/jul/10/us-navy- drone-swarm-in-action/.
x47b-drone-aircraft-carrier. 61 Mark Prigg, “Pentagon to Use AI to Defeat ISIS in Iraq and Syria |
58 “X-47B UCAS Makes Aviation History…Again!,” Northrop Grum- Daily Mail Online,” Daily Mail, May 15, 2017, https://www.dailymail.
man, April 22, 2015, https://www.northropgrumman.com/Capabilities/ co.uk/sciencetech/article-4508874/Pentagon-use-AI-defeat-ISIS.html.
X47BUCAS/Pages/default.aspx. 62 Andrew Liptak, “The US Air Force’s Jet-Powered Robotic Wingman Is
59 Chris Baraniuk, “US Military Tests Swarm of Mini-Drones Launched like Something out of a Video Game,” The Verge, March 9, 2019, https://
from Jets,” BBC, January 10, 2017, https://www.bbc.com/news/technol- www.theverge.com/2019/3/9/18255358/us-air-force-xq58-a-valkyrie-
ogy-38569027. prototype-robotic-loyal-wingman-drone-successful-test-flight.

s e ta v. o rg 17
 ANALYSIS

tests have been conducted by the Army and the Xi Jinping, on many occasions has clearly stated
Marine where prototype vehicles have followed that the leadership in AI technology is “critical to
and protected soldiers or vehicles in the battle- the future of global military and economic power
field. Sea Hunter is another prototype vessel (Anti- competition.”65 It can be said that for China, the
Submarine Warfare Continuous Trail Unmanned military AI R&D is seen as a possible and easy way
Vessel prototype) which was first developed by to challenge the American military hegemony.
DARPA to be later transferred to the Office of The Chinese government has increased
Naval Research. Sea Hunter is the first drone ship their AI R&D spending by 350% between 2005
which has traveled autonomously from California and 2015 and it is considered to be very close to
to Hawaii and then back to California.63 the U.S. currently. Furthermore, in 2017, 48%
of the world’s total AI start-up funding was cov-
It can be said that for China, ered by Chinese companies and between 2013
the military AI R&D is seen as a and 2018 China’s AI industry attracted 60% of
global funding for AI.66 As a result, it is believed
possible and easy way to challenge
among the Chinese leadership and industry that
the American military hegemony. the gap between China and the U.S. in AI is very
Lastly, the incorporation of AI into LAWS narrow now and China sees “AI as ‘a race of two
remains of crucial importance. However, the giants,’ between itself and the United States.”67
U.S. has not declared any tests or experiments in While there may be a lot of skepticism re-
this regard until now. garding China’s challenge to the American hege-
mony, it can be said that the systemic and strategic
advantages can act in favor of China and help it
CHINA AND ARTIFICIAL achieve its goal of becoming a world leader in AI.
INTELLIGENCE This would include the potential human talent re-
sources and the large amount of data that China
China is following in the steps of the U.S. when
possesses but most importantly the relation of the
it comes to the research and application of AI
government and AI private sector.68 In this context,
technology. Not only is it using AI in domestic
in contrast to the U.S., one of China’s strongest
surveillance, but China has already stated that it
strategies is the civil–military integration69 (CMI)
aims to overtake the West in AI R&D by 2025
and the development of advanced dual-use tech-
and more importantly to be the world leader in AI
by 2030.64 The Chinese leadership, including here
65 Gregory C. Allen, “Understanding China’s AI Strategy: Clues to Chi-
nese Strategic Thinking on Artificial Intelligence and National Security”
63 Joseph Trevithick, “Navy’s Sea Hunter Drone Ship Has Sailed Au- (Washington, DC: Center for a New American Security, 2019).
tonomously To Hawaii And Back Amid Talk Of New Roles,” The Drive, 66 Meia Nouwens, “China’s Pursuit of Advanced Dual-Use Tech-
April 2, 2019, https://www.thedrive.com/the-war-zone/26319/usns-sea- nologies,” IISS, December 18, 2018, https://www.iiss.org/blogs/analy-
hunter-drone-ship-has-sailed-autonomously-to-hawaii-and-back-amid- sis/2018/12/emerging-technology-dominance.
talk-of-new-roles.
67 Allen, “Understanding China’s AI Strategy: Clues to Chinese Strategic
64 Pablo Robles, “China Plans to Be a World Leader in Artificial Intel- Thinking on Artificial Intelligence and National Security,” 9.
ligence by 2030,” South China Morning Post, October 1, 2018, https://
multimedia.scmp.com/news/china/article/2166148/china-2025-artifi- 68 Baidu, Alibaba and Tencent (called also as BAT) are the main compa-
cial-intelligence/index.html; Cade Metz, “As China Marches Forward on nies that lead the Chinese AI industry.
A.I., the White House Is Silent,” The New York Times, February 12, 2018, 69 Civil-military integration (or sometimes even called as civil-military
sec. Technology, https://www.nytimes.com/2018/02/12/technology/ fusion) means that the technologies which are civilian in application can
china-trump-artificial-intelligence.html. be used also for defense purposes. AI is an example of these technologies.

18 s e ta v. o rg
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

nologies.70 These complementary strategies are re- • 13th Five Year Plan for Developing National
lated to the fact that the Chinese government and Strategic and Emerging Industries (2016-
AI private sector are working closely together and 2020) (August 2016): Based on this plan,
this makes the application of AI technologies to the the State Council contended that AI was
military easier. The main goal behind such a policy among the main tasks to be pursued by
is to create a strong military and help the People’s the central government (6 out of 69 major
Liberation Army (PLA) to dominate the warfare tasks) and at the same time it announced
domains and as a result ‘leapfrog’ the U.S.71 five agencies that would be responsible for
One turning point for China’s perception the development of AI.74
and stance towards the AI R&D was in 2016 • Next Generation Artificial Intelligence De-
when Google DeepMind’s AlphaGo beat the velopment Plan (July 2017): This can be
world champion Lee Sedol by 4-1 in a Go game. considered as one of the main documents
Many have called this the Sputnik moment for regarding the Chinese AI strategy. This is
China as the government started to put more the plan in which China explicitly declares
importance on the AI national strategy and its the goal to become a world leader in AI by
implementation in the military. Unlike the U.S. 2030. Furthermore, the plan identifies a
the AI national strategies in China are released specific path that needs to be followed in
mainly by China’s State Council. The plans and order to achieve the proposed goal. The
programs released can be summarized as below: document argues that AI embodies a “ma-
• Made in China 2025 (May 2015): This jor strategic opportunity” and suggests a
10-year action plan focused on promoting coordinated strategy to “build China’s first
manufacturing, among others in the field of mover advantage.”75
artificial intelligence and robotics.72 • 13th Five-Year Science and Technology
• “Internet Plus” Artificial Intelligence Three- Military-Civil Fusion Special Plan (Au-
Year Action Plan (2016-2018) (May 2016): gust 2017): In contrast to the others, this
This plan was focused specifically on AI and plan was issued by the Ministry of Science
called for the creation of infrastructure and and Technology and the CMC Science
platforms for AI. Furthermore, it estab- and Technology Commission. As the title
lished a goal to increase the AI industry to suggests this plan focuses directly on the
a level totaling billions of RMB by 2018.73 military-civil fusion and the dual use of AI
technology and reflects on the importance
70 Meia Nouwens, “China’s Pursuit of Advanced Dual-Use Technologies.”
of this strategy.76
71 Elsa B. Kania, “Battlefield Singularity: Artificial Intelligence, Military
Revolution, and China’s Future Military Power” (Washington, DC: Cen-
ter for a New American Security, 2017); Carter, Kinnucan, and Elliot, “A 74 Compilation and Translation Bureau, trans., “13th Five Year Plan for
National Machine Intelligence Strategy or the United States.” Developing National Strategic and Emerging Industries (2016-2020)”
72 “‘Made in China 2025’ Plan Issued,” The State Council, May 19, (Central Committee of the Communist Party of China, 2016), http://
2015, http://english.gov.cn/policies/latest_releases/2015/05/19/con- en.ndrc.gov.cn/newsrelease/201612/P020161207645765233498.pdf.
tent_281475110703534.htm; James McBride and Andrew Chatzky, “Is 75 China’s State Council, “Full Translation: New Generation Artificial
‘Made in China 2025’ a Threat to Global Trade?,” Council on Foreign Intelligence Development Plan (2017),” trans. Graham Webster et al.,
Relations, March 7, 2019, https://www.cfr.org/backgrounder/made-china- New America, accessed April 23, 2019, https://www.newamerica.org/
2025-threat-global-trade. cybersecurity-initiative/digichina/blog/full-translation-chinas-new-gen-
73 Xiaomin Zhao, “Development Strategy Analysis of ‘Internet Plus’ Ar- eration-artificial-intelligence-development-plan-2017/.
tificial Intelligence Technology,” Advances in Intelligent Systems Research 76 Kania, “Battlefield Singularity: Artificial Intelligence, Military Revo-
147 (2018). lution, and China’s Future Military Power.”

s e ta v. o rg 19
ANALYSIS

• Robotic Industry Development Plan (2016- Military Commission; while the first one is
2020) (April 2016): was issued by three in- mainly responsible for the civilian application
stitutions, namely National Development of technology, the latter is responsible for its
and Reform Commission (NDRC), Minis- military applications. This division also reflects
try of Industry and Information Technolo- China’s civil-military integration national strat-
gy (MIIT) and Ministry of Finance (MOF). egy mentioned previously.
The importance of this plan rests on the fact The Leading Small Group is composed of
that it set concrete goals and strategies in the main figures of the State Council -China’s
terms of the robotics industry. More specifi- top administrative body- and is led by the leader
cally, this plan states that China is planning of the State Council, the Chinese Premier. Based
to produce 100,000 robots by 2020.77 on the 13th Five-Year Plan for Developing Na-
• Three Year Action Plan for Promoting Devel- tional Strategic and Emerging Industries there are
opment of a New Generation Artificial In- five main institutions responsible for the develop-
telligence Industry (2018-2020) (December ment of AI: National Development and Reform
2017): This plan was issued by the Chinese Commission (NDRC), State Internet Informa-
Ministry of Industry and Information Tech- tion Office (SIIO), Ministry of Industry and
nology and provides further details on those Information Technology (MIIT), Ministry of
published in the Next Generation Artificial Science and Technology (MOST) and Ministry
Intelligence Development Plan. This plan of Finance (MOF). Nevertheless, there are other
identifies the way to stimulate the develop- institutions, which fall under the State Council,
ment of AI in the 2018-2020 period and also that are also responsible for the AI R&D. These
the key areas of AI application, including ve- may include the Ministry of Education, the Chi-
hicles, service robots, drones, etc.78 nese Academy of Sciences (CAS), the Chinese
Academy of Engineering (CAE), the National
Natural Science Foundation of China (NSFC).
Bureaucratic Structure On the other hand, the military aspect of
Considering that China is a one-party state, the
technology development is under the control
Chinese Communist Party Central Committee
of the Central Military Commission – which
controls all the activities including the techno-
commands and controls the People’s Libera-
logical developments. The policy making and
tion Army (PLA), the People’s Armed Police and
coordination regarding technology is conducted
the Militia. Within this framework, AI R&D is
through two main institutions: National Sci-
carried out mainly by military universities and
ence, Technology and Education Leading Small
research institutes. The PLA, which is the main
Group and Chinese Communist Party Central
agency in terms of AI R&D, expects AI to have a
77 Yujia He, “How China Is Preparing for an AI-Powered Future” fundamental impact on the character of warfare.
(Washington, DC: Wilson Center, June 2017), https://www.wilson- Accordingly, we will see a shift from the ‘infor-
center.org/sites/default/files/how_china_is_preparing_for_ai_pow-
ered_future_0.pdf. matized warfare’ that is currently applied to the
78 Ministry of Industry and Information Technology, “Translation: ‘intelligentized warfare.’79 Furthermore, China’s
Three Year Action Plan for Promoting Development of a New Genera-
tion Artificial Intelligence Industry (2018-2020),” trans. Paul Triolo, Elsa
B. Kania, and Graham Webster, New America, January 2018, https://
www.newamerica.org/cybersecurity-initiative/digichina/blog/transla- 79 Kania, “Battlefield Singularity: Artificial Intelligence, Military Revo-
tion-chinese-government-outlines-ai-ambitions-through-2020/. lution, and China’s Future Military Power,” 5.

20 s e ta v. o rg
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

main focus remains in the application of AI in the The war-gaming is seen from the PLA as a
decision-making, military deduction, unmanned way to help it in the application of AI in mili-
military weapons, and defense equipment; yet, it tary applications which would help China win
is possible to say that the Chinese military appli- in future warfare. However, this is only a small
cation of AI is at its early stage. The main reason part of the AI military applications that China
for this is that, unlike the U.S., China does not has achieved until now. The lack of real combat
have actual experience in the battlefield. So, the experience has not stopped China from research-
main focus of the PLA appears to be the applica- ing and testing the application of AI in autono-
tion of the AI in simulations and war-gaming. mous unmanned systems, its use for defense and
offense, and also the support of the command
AI Military Applications in China decision-making by AI.
Even though it is dissimilar from the U.S., China
is taking considerable steps regarding the incor- Focusing on the autonomous
poration of AI into its defense systems, mainly unmanned systems, similar
the military. In 2015, Baidu – one of China’s to the U.S., China has been
leading AI companies – created a language rec-
researching air, ground, surface
ognition software that surpassed the human
levels. It is important to emphasize the fact that
and undersea autonomous
Microsoft was not able to create a similar soft- unmanned vehicles (AUV).
ware until a year later. The language recognition
software, accompanied by the computer vision Focusing on the autonomous unmanned
systems that are developed in China are currently systems, similar to the U.S., China has been re-
being used for domestic surveillance in order to searching air, ground, surface and undersea au-
monitor civilians. tonomous unmanned vehicles (AUV). In terms
If we would focus on the military applica- of the air AUVs, China has been quite success-
tion of AI in China, it is important to mention ful especially when it comes to swarm drones.
that because of the lack of combat experience, In June 2017, China managed to fly a swarm of
China is more focused on war-games and simu- 119 drones – so breaking the American record
lations on the basis of augmented reality.80 One of 103 drones- all of which were equipped with
of the most well-known examples is the Artifi- systems that allowed the drones to communicate
cial Intelligence and War-Gaming National Fi- with each other.81
nals organized in 2017 by the China Institute TYW-1 and ASN-216 are two examples
of Command and Control. In this competition of the Chinese unmanned aerial vehicle (UAV)
humans were confronted with machines which which with the help of AI can now operate au-
acted without any human control, known as tonomously, though not fully. Currently, both
CASIA-Prophet 1.0. The result of this competi- ASN-216 and TYW-1 can take off and land
tion was 7 to 1 in favor of the machine. without human intervention and TYW-1 with

80 Elsa Kania, “Learning Without Fighting: New Developments in PLA

Artificial Intelligence War-Gaming,” China Brief 19, no. 7 (April 9, 81 Emily Feng and Charles Clover, “Drone Swarms vs Conventional
2019), https://jamestown.org/program/learning-without-fighting-new- Arms: China’s Military Debate,” Financial Times, August 24, 2017,
developments-in-pla-artificial-intelligence-war-gaming/. https://www.ft.com/content/302fc14a-66ef-11e7-8526-7b38dcaef614.

s e ta v. o rg 21
ANALYSIS

a minimal human intervention can even identify The outcome of the U.S.-China competi-
and attack a target.82 tion is not clear as yet; however, it can be said
When it comes to the unmanned surface ve- that the threat posed to the U.S. by China should
hicles, SeaFly is an example that needs to be men- not be underestimated. There is a possibility that
tioned. Currently, SeaFly can learn how to avoid the U.S. keeps the advantage for now, but in
obstacles without human intervention and at the the long run China possesses great capabilities
same time it is able to recover UAV by using al- to challenge this position. As mentioned above,
gorithms that allow SeaFly to arrange its actions the U.S. announced it national strategy approxi-
on the basis of its assessment of sea conditions.83 mately one year after China did. Yet many ex-
Lastly, in terms of the implementation of AI perts have argued that detailed information of
into LAWS, in 2018, China’s delegation to the funding is lacking in this strategy.
UN Group of Governmental Experts on LAWS At this point, it is crucial that the U.S. does
contended that they accept the conclusion of a not take its military superiority for granted and
new protocol based on which the usage of fully belittle the steps undertaken by China. More
autonomous lethal weapons systems would be specifically, while clearly the U.S. was superior
banned; however, for China, this does not mean in the Second Offset strategy which was focused
that the development of fully autonomous lethal on the precision-guided weapons, the same po-
weapons systems would be banned.84 sition may not be preserved with the Third Off-
set strategy, of which AI is a part. The reasons
for this are many.
CONCLUSION First of all, unlike the U.S., China applies a
The capacity of AI to alter the landscape of se- ‘whole-of-government’ approach, meaning that
curity and change the current balance of power it applies a top-down process where the govern-
is obvious and accepted by both bureaucrats ment plays a vital role and has a direct impact on
and academicians. For this reason, the rise in AI R&D. This allows the Chinese government
the development of AI, and its application in to control the process more easily. Furthermore,
the military, has been accompanied by a com- the main private companies responsible for AI
petition between states, most notably the U.S. R&D, Baidu, Alibaba and Tencent, have already
and China. Many other states at the same time announced that they will work closely with the
are trying not to lag behind and advance their Chinese government in terms of AI R&D. How-
own AI applications. ever, the same thing cannot be said for the U.S.
As the U.S. government does not have direct
82 Kania, “Battlefield Singularity: Artificial Intelligence, Military Revo- control over the companies responsible for AI
lution, and China’s Future Military Power,” 22.
R&D, i.e. Google, Microsoft, etc., it makes it
83 Mark Episkopos, “Underwater Stealth Swarms?: Introducing China’s
SeaFly Unmanned Stealth Vessel,” Text, The National Interest, Decem- more difficult to advance AI for the U.S. An ex-
ber 26, 2018, https://nationalinterest.org/blog/buzz/underwater-stealth- ample of this is the case of Project Maven, which
swarms-introducing-chinas-seafly-unmanned-stealth-vessel-39882.
84 Elsa Kania, “China’s Strategic Ambiguity and Shifting Approach to
was a collaboration between the Department of
Lethal Autonomous Weapons Systems,” Lawfare, April 17, 2018, https:// Defense and Google. But the latter had to with-
www.lawfareblog.com/chinas-strategic-ambiguity-and-shifting-ap-
proach-lethal-autonomous-weapons-systems; Kirsten Gronlund, “State draw from the project as thousands of its em-
of AI: Artificial Intelligence, the Military and Increasingly Autonomous ployees signed a petition to end the use of their
Weapons,” Future of Life Institute, May 9, 2019, https://futureoflife.
org/2019/05/09/state-of-ai/. work by the military.

22 s e ta v. o rg
 ARTIFICIAL INTELLIGENCE APPLICATION IN THE MILITARY: THE CASE OF UNITED STATES AND CHINA

Secondly, China possesses robust human The outcome of the U.S.-China

capital potential and a massive data resource base competition is not clear as yet;
which could help them gain the upper-hand in
however, it can be said that the
AI R&D. It is estimated that China is on track to
possess 20% of the world’s share of data by 2020, threat posed to the U.S.
with the potential to have over 30% by 2030.85 by China should not be
Furthermore, China is more prone, than the underestimated.
U.S., to not take into consideration the human-
the few states that has developed a national UAV,
factor challenges associated with full autonomy
it is crucial that it advances in AI R&D and its
and the application of AI in LAWS. While the
application in the military sector. There is an
United States has clearly announced that it does
urgent need for an AI national strategy and the
not intend to keep humans out of the loop, the
reasons for this are twofold. First of all, the risk
same thing does not stand for China. If China
of lagging behind in the AI competition may
succeeds in applying full autonomy and applies
prove detrimental for Turkey. The application of
AI in LAWS this would result in a warfare asym-
AI in the military provides first-movers advan-
metry in China’s favor.
tages and those states that do not apply the AI in
Yet, there is a drawback for China in terms
the short run, will face greater threats in the long
of battlefield experience. Considering that China
run when it will be very difficult to keep-up with
has not been in a war in recent decades it is dif-
the pace. As a result, the successes that Turkey
ficult for it to apply the AI directly in the battle-
has achieved until now in terms of national de-
field. For this reason, China has been focusing
fense, especially regarding the UAV, will be un-
mainly on war games and simulations. However,
dermined considering that other regional states,
the U.S. which has continuously been at war has
such as UAE have already taken important steps
been more concentrated on more tactical-level
in terms of AI R&D. Secondly, in recent years,
application and as such it has the advantage of
Turkey has been surrounded by different threats,
applying AI in the battlefield and so developing
including here PKK, YPG, ISIS, etc. AI is very
it according to the needs of combat.
difficult to develop, however, once the software is
The U.S. and China are not the only two
created copying and applying it is very easy. This
states that have turned AI into a policy priority.
means that it would not be difficult for terror-
As mentioned earlier, there are at least 20 states
ist organizations, such as PKK, YPG or ISIS, to
that have already announced their AI national
use AI in order to achieve their goals and at the
strategies and most of them have undertaken im-
same time threaten the national security of Tur-
portant steps in AI R&D. Yet, Turkey has neither
key. However, if Turkey starts to focus on AI and
announced any AI national strategy nor taken
make it a policy priority, as many other states
steps in AI R&D or its application in the mili-
have done, it would be easier for the country to
tary. Considering the fact that Turkey is among
protect itself from any future threat that would
come from the possession of AI technology by
85 Kania, “Battlefield Singularity: Artificial Intelligence, Military Revo-
lution, and China’s Future Military Power,” 12. such terrorist organizations.

s e ta v. o rg 23
 ARTIFICIAL INTELLIGENCE
APPLICATION IN THE MILITARY:
THE CASE OF UNITED STATES AND CHINA
GLORIA SHKURTI ÖZDEMIR

ANALYSIS

Considered as the 4th Industrial Revolution, Artificial Intelligence (AI) has

become a reality in today’s world, especially in the military. Experts and aca-
demicians have emphasized the importance of AI for a long time. Further-
more, world leaders, including Obama, Trump, Xi, and Putin, have all made
important statements that bring to the fore the significance of AI which can
be summarized with what Putin stated on September 2017: whoever beco-
mes the leader in AI, will rule the world.
This analysis provides a short introduction on what AI is, how it has evolved
until today and how it will change the nature of warfare. It then assesses
why states invest in AI to later turn to the case of the U.S. and China. For
both states, the main official documents and statements are analyzed, the
bureaucratic structures that work on AI are presented and finally examples
of how the U.S. and China are applying AI in the military are provided. The
conclusion briefly comments on how the strategies of China and the U.S. dif-
fer followed by some recommendation on what states like Turkey should do
in the near future.

www.setav.org

ANKARA • İSTANBUL • WASHINGTON D.C. • KAHİRE • BERLİN

View publication stats