Hague Centre for Strategic Studies

Report Part Title: Domain impacts

Report Title: Macro Implications of Micro Transformations

Report Subtitle: An Assessment of AI’s Impact on Contemporary Geopolitics
Report Author(s): Hugo van Manen, Salma Atalla, Amit Arkhipov-Goyal and Tim Sweijs
Published by: Hague Centre for Strategic Studies (2019)
Stable URL: https://www.jstor.org/stable/resrep19557.6

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Hague Centre for Strategic Studies is collaborating with JSTOR to digitize, preserve and
extend access to this content.

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 2. Domain impacts

The included case studies’ initiatives facilitate their active engagement with and
development of AI-related technologies whose emergence impacts the economic,
sociopolitical, and military-security domains. AI’s application within these domains
can be associated with a ream of ‘generic’ impacts which are shortly outlined in the
sections below, which include both threats and opportunities. Within the context of
this study, ‘generic’ impacts are derived from a ‘horizon scan’-style literature review of
±50 documents published by research institutes, think tanks, NGOs & consultancies,
and academia, while more concrete (domain-specific) implications are derived from
observations made from the case studies themselves. More specifically, we identify
concrete negative externalities which are associated with initiatives undertaken
by actors within the study. This approach is underpinned by several overarching
assumptions; namely:

a) that current activities are indicative of future activities,

b) that more productive ecosystems are better equipped to push the envelope than
those which are not, and
c) that a higher degree of strategic coherence amplifies ecosystem productivity
and output.

This means that – in general – the concrete negative externalities outlined in the
sections below constitute extrapolations based on past and present initiatives.

2.1 Economic
The economic impacts of AI-related technologies are acutely evident in society as
a result of the private sector role in the development and use of AI for commercial
applications. Prospect of gains have prompted companies to seek applications of AI to
outperform competitors, both nationally and internationally.112 While the processing
of Big Data has provided ample opportunities - namely: supply-side efficiencies,
personalization of goods and services and stimulation of the economy - the benefits

112 Directorate-General for Research and Innovation, “2018 Industrial R&D Scoreboard: EU Companies Increase
Research Investment amidst a Global Technological Race” (European Commission, December 17, 2018),
https://ec.europa.eu/info/news/2018-industrial-rd-scoreboard-eu-companies-increase-research-investment-
amidst-global-technological-race-2018-dec-17_en.

54 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
of these are being distributed unevenly. AI’s effectiveness depends heavily on the
quantity and quality of data, elements often reserved to the leading corporations
in the digital sector.113 Data generation and its use is creating path dependencies,
as companies with the greatest data accumulation continuously reinforce their
leadership position and stifle market competition.114 The adverse effects of these
trends are evident in modern markets’ tendency towards developing oligopolistic
markets with winner-takes-all-dynamics, the rise in inequality, economic insecurity
and a protectionist response with the creation of rival regulatory frameworks.
Moreover, these threats are becoming evident at the individual, market and state
levels. A high-level overview of the threats and opportunities associated with AI’s
application within the economic domain is provided in Figure 7 below.

Threats Opportunities
Winner-takes-all dynamics 
Emergence of oligopolistic market structures   Improves supply-side efficiency
Manifests AI ‘haves’ and ‘have-nots’   Allows for personalization of goods and services
Leads to protectionist regulatory frameworks   Facilitates market and trade expansion
Impacts economic security 

Economic

Figure 7: Economic Threats and Opportunities of AI.

2.1.1 Threats

Digitization of processes across many sectors in the economy, namely finance,

healthcare, manufacturing, logistics, and consumer goods and services, has given rise
to a new dimension of competition. Companies that harvest mass amounts of data

113 Willem Sundblad, “Data Is The Foundation For Artificial Intelligence And Machine Learning,” Forbes, October
18, 2018, https://www.forbes.com/sites/willemsundbladeurope/2018/10/18/data-is-the-foundation-for-artificial-
intelligence-and-machine-learning/.
114 Daniel Faggella, “The AI Advantage of the Tech Giants: Amazon, Facebook, and Google,” Emerj, February 26,
2019, https://emerj.com/ai-executive-guides/ai-advantage-tech-giants-amazon-facebook-google/.

Macro Implications of Micro Transformations 55

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 - oftentimes generated by their processes or customers - leverage it to improve their
competitive position.115 By utilizing the accumulated Big Data, companies accelerate
the development of AI technologies. This dynamic positions companies with the
largest data generation potential to benefit disproportionately from AI. As a result,
integration of AI is contributing to the development of winner-takes-all-dynamics
in (inter)national markets. This effect is particularly evident in the technology sector,
where network effects continue to generate disproportionately large returns for the
leading firms such as Google and Facebook, with the result being the development of
oligopolistic market structures and inequality.116

The winner-takes-all-dynamics demonstrates how AI is reinforcing the tendency

towards oligopolistic market structures. This occurs, in part, from the cost of
developing and integrating AI-related technologies into industrial workflows
and the uneven implementation of such technologies within industrial sectors.117
Furthermore, due to implementation costs, industrialized economies are likely to
realize the efficiencies associated with the technology before developing ones.118 As
an outcome, the winner-takes-all-dynamics reinforce oligopolistic market structures,
within which an increasingly small number of corporations dominate (inter)national
markets.119 This dynamic is particularly likely to propagate geopolitically negative
externalities within sectors such as agriculture and manufacturing, where efficiency
gains of Western and/or Chinese firms are likely to render them relatively more
competitive to their counterparts in developing nations. These changes are set to
reduce the prosperity of producers in developing countries, which will be less able
to compete with their Western counterparts on price and quality, thus reducing a
significant share of these states’ income potential. This dynamic also plays out at the
domestic level, where often already marginalized population groups will experience
the negative externalities associated with these technologies’ implementation. Across
the developed nations, blue-collar workers will be affected disproportionately more
than their white-collar counterparts.

AI’s impact on inequality is widespread, with AI haves and have-nots forming at

individual, market and state levels. Aside from the effect on traditional sectors,

115 Faggella.
116 “AI, China, Russia, and the Global Order: Technological, Political, Global, and Creative Perspectives”
(Washington, D.C.: Department of Defense, 2018), 38, https://nsiteam.com/social/wp-content/uploads/2018/12/
AI-China-Russia-Global-WP_FINAL.pdf.
117 Jacques Bughin et al., “Notes from the AI Frontier: Modeling the Global Economic Impact of AI” (McKinsey),
accessed April 13, 2019, https://www.mckinsey.com/~/media/McKinsey/Featured%20Insights/Artificial%20
Intelligence/Notes%20from%20the%20frontier%20Modeling%20the%20impact%20of%20AI%20on%20the%20
world%20economy/MGI-Notes-from-the-AI-frontier-Modeling-the-impact-of-AI-on-the-world-economy-
September-2018.ashx.
118 PricewaterhouseCoopers, “PwC’s Global Artificial Intelligence Study: Sizing the Prize,” PwC, accessed June 28,
2019, https://www.pwc.com/gx/en/issues/data-and-analytics/publications/artificial-intelligence-study.html.
119 Arif Khan, “Disrupt the Disruption: The Tech Oligopoly Part 2,” SingularityNET, October 7, 2018, https://blog.
singularitynet.io/disrupt-the-disruption-the-tech-oligopoly-part-2-bb8747b7e16d.

56 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
tech clusters and the accompanying high-paying salaries are concentrated in a
number of developed countries, indicating a tech talent flow to these clusters and
resulting in a harder hit on developing nations’ economies.120 Within nation states,
the rural-urban divide has been highlighted in recent years by the rise in populism,
where the disenfranchised rural areas continue to stagnate while the largest cities
flourish.121 While AI is not responsible for this trend, it is acting as a stress factor on
the current socio-economic trends. Within industries, AI is acutely impacting low-
skilled professions, where increases in automation are resulting in human tasks being
replaced or even becoming obsolete. Within the technology sector, technical jobs
are male-dominated, particularly in countries leading the AI development race. This
sectoral gender-divide highlights the risks that gender inequality will be ingrained in
AI due to bias in the development process.122 This is further leading to the unequal
distribution of benefits, with few policy suggestions currently in place to offset this.

AI is perceived to further establish ‘AI haves and have-nots’ among nation states. It
threatens to eliminate comparative advantages of developing economies that rely
heavily on cheap, labor-intensive manufacturing, such as textile production.123 While
countries such as China and the US are predicted to be the top beneficiaries by taking
70% of AI’s economic benefits, developing countries across Asia, Africa and Latin
America are expected to gain less than 6% of the prospected global GDP increase.124
The overall result is anticipated wider income, gender, geographic, and international
inequalities.125 It is notable, that the extent of these technologies’ overall impact on
the labor market is currently inconclusive, with evidence of both positive and negative
trends, whereby while some jobs are displaced, new opportunities are created. Past
waves of technological progress have often resulted in worker displacement, while at
the same time generated net positive gains in employment. However, estimates for
the current technological revolution estimate the impact to take effect at ten times
the pace of the previous ones.126 As a result, it is evident that to harness the benefit
of AI, policies have to be adopted to accommodate transition of displaced workers to
other professions.

120 “AI, China, Russia, and the Global Order: Technological, Political, Global, and Creative Perspectives.”
121 Gideon Rachman, “Urban-Rural Splits Have Become the Great Global Divider,” Financial Times, July 30, 2018,
https://www.ft.com/content/e05cde76-93d6-11e8-b747-fb1e803ee64e.
122 Hannah Devlin and Alex Hern, “Why Are There so Few Women in Tech? The Truth behind the Google
Memo,” The Guardian, August 8, 2017, sec. Life and style, https://www.theguardian.com/lifeandstyle/2017/
aug/08/why-are-there-so-few-women-in-tech-the-truth-behind-the-google-memo; John Villasenor, “Artificial
Intelligence and Bias: Four Key Challenges,” Brookings (blog), January 3, 2019, https://www.brookings.edu/blog/
techtank/2019/01/03/artificial-intelligence-and-bias-four-key-challenges/.
123 “AI, China, Russia, and the Global Order: Technological, Political, Global, and Creative Perspectives.”
124 PricewaterhouseCoopers, “PwC’s Global Artificial Intelligence Study.”
125 Though AI is likely to result in job loss in the short term, the technology’s implementation will also create jobs
and/or greatly increase demand for several existing professions. See McKinsey & Company, Inc., “Smartening up
with Artificial Intelligence (AI) - What’s in It for Germany and Its Industrial Sector?,” April 2017, https://www.
mckinsey.com/~/media/McKinsey/Industries/Semiconductors/Our%20Insights/Smartening%20up%20with%20
artificial%20intelligence/Smartening-up-with-artificial-intelligence.ashx.
126 Klaus Schwab, “Globalization 4.0,” January 22, 2019, 0, https://www.foreignaffairs.com/articles/
world/2019-01-16/globalization-40.

Macro Implications of Micro Transformations 57

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 Exacerbation of inequalities vis-à-vis the rate of harvesting the benefits of AI may
result in laggards initiating protectionist regulatory frameworks to resist the
adoption of rules which they view as ‘adding an extra burden on them, with vague
benefits’ within international institutions.127 An intensification of protectionism is
likely to follow as some states are incentivized to introduce measures geared towards
protecting domestic industries from international competition. This fear of ‘losing
out’ incentivizes countries to hamper the development of international standards
and introduce barriers to international trade. The resistance is reinforced by the
states perception that conceding to these rules risks the inception of an international
framework which consolidates existing (digital) power asymmetries.128 This may
hamper the timely adoption of internationally binding norms and rules, thus
potentially exacerbating the severity of existing digital divides. Besides domestic
protectionist measures such as the US’ assault on the Chinese technology giant
Huawei, barriers are being erected at multilateral institutions. In light of the rising
trade tensions between the US and China, the US has actively slowed down the
functioning of the dispute-settling Appellate Body at the World Trade Organization
(WTO) by failing to reappoint judges and thus reducing their ability to handle cases in
a timely manner.129

Introduction of AI is further presenting challenges of economic insecurity, both at

domestic and international levels. At the domestic level, the prominence of the private
sector in the development and implementation of AI increases the vulnerability
of governments and alters their strategic options. This is a result of control over
crucial technologies, often supporting critical infrastructure, being handed over
to a third-party. This has national security connotations, as highlighted by, among
others, the US presidential order blocking the acquisition of the American chip
manufacturer Qualcomm by Singaporean Broadcom.130 At the international level,
recent challenges have been safeguarding companies from corporate espionage and
predatory business practices abroad, particularly in China. Companies from Western
market-based economies find themselves competing against (Chinese) companies
or state-owned enterprises with substantial state financial backing, which results in
market distortions and hampers the innovation and growth of Western companies.
The arising deadlock in the global regulatory environment and the slowdown of
dispute settlement at the WTO negatively affects economic security as it limits the

127 Christopher Foster and Shamel Azmeh, “Trade Wars Are Growing over the Digital Economy – and Developing
Countries Are Shaping the Agenda,” The Conversation, 2019, http://theconversation.com/trade-wars-are-
growing-over-the-digital-economy-and-developing-countries-are-shaping-the-agenda-113000.
128 This view is further exacerbated by the fact that many developing countries preside over neither the technical
know-how nor the foreknowledge to predict how their domestic industries will develop and are therefore unable
to comprehensively evaluate these rules’ likely impact.
129 “U.S. Blocks WTO Judge Reappointment as Dispute Settlement Crisis Looms,” Reuters, August 27, 2018,
https://www.reuters.com/article/us-usa-trade-wto-idUSKCN1LC19O.
130 Jacob Kastrenakes, “Trump Issues Order Blocking Broadcom Takeover of Qualcomm, Citing National Security,”
The Verge, March 12, 2018, https://www.theverge.com/2018/3/12/17111766/broadcom-qualcomm-acquisition-
blocked-trump-national-security.

58 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
international potential of companies. A further issue is corporate espionage, which
has become more prolific in the age of digitization. Leading Western technology
companies are consistently under attack from groups seeking to acquire commercial
technology.131 Reinforced by predatory business practices, market players are
challenged further through unfair intellectual property (IP) transfers, whether legal
or illegal. In the case of China, most companies simply handed over their IP as a
prerequisite to gain access to the Chinese market.132 The impact is significant, as in
numerous instances, technology produced by Western companies is being used to
reinforce digital totalitarianism in authoritarian states.133 Due to the aforementioned
dependence on the private sector in digitization, corporate espionage threatens to
undermine not only economic, but also national security.

Overall, the economic threats arising from AI are ample and will have wide-reaching
implications. Within markets, introduction of AI risks to reinforce the position of
the leading firms in a winner-takes-all-dynamic, stifle competition and thus limit
innovation. The uneven distribution of AI and automation within the economy will
likely result in increases in inequality across numerous boundaries, between, among
others, the digitally literate vs illiterate, blue-collar vs white-collar workers and
rural vs urban residents. The resulting adverse effects are prompting governments
to establish protectionist measures to safeguard domestic industries and limit the
negative externalities of the exponentially growing impact of AI technologies. Finally,
the economic risks involving the companies behind AI development have implications
on national security by limiting the strategic position of states, as well as exposing
critical industries to foreign predatory practices and corporate espionage.

2.1.2 Opportunities

AI-related technologies are projected to have substantial ramifications on domestic,

regional and global economies, with figures estimating a global contribution in GDP
of as much as $13 trillion by 2030, equal to 1.2 percent of additional GDP growth per
year.134 AI’s added value within the economic domain currently derives almost entirely
from the technology’s ability to automate processes. Applications span manufacturing
tasks, particularly those in unsafe environments, quality testing, logistics, fraud
detection, and routine business processes. It can be universally associated with

131 “China Broke Hacking Pact before New Tariff Fight,” Axios, accessed June 28, 2019, https://www.axios.com/
china-broke-hacking-pact-before-new-tariff-tiff-d19f5604-f9ce-458a-a50a-2f906c8f12ab.html.
132 United States International Trade Commission, “China: Effects of Intellectual Property Infringement and
Indigenous Innovation Policies on the U.S. Economy” (US Government, 2011), https://www.usitc.gov/
publications/332/pub4226.pdf.
133 Sui-Lee Wee, “China Uses DNA to Track Its People, With the Help of American Expertise,” The New York Times,
February 21, 2019, sec. Business, https://www.nytimes.com/2019/02/21/business/china-xinjiang-uighur-dna-
thermo-fisher.html.
134 Bughin et al., “Notes from the AI Frontier: Modeling the Global Economic Impact of AI.”

Macro Implications of Micro Transformations 59

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 increased safety, cost efficiency, productivity, empowerment of workers and advanced
understanding of market demand.

AI allows for unprecedented gains in supply-side efficiency, driven by improved

worker productivity, cost-efficiency and process optimization. AI is reducing the
necessity for humans to operate in dangerous environments such as offshore oil rigs
and coal mines, ensuring the quality of products through image recognition, and
reducing oversupply and loss of sales through automating stock monitoring and
replenishment. Furthermore, it is automating routine business processes in data
management, provision of IT services, and semantic analysis. Particular aspects of
supply chains that will benefit from AI are repetitive tasks, processes generating
large amounts of data and work requiring pattern or anomaly detection from text or
imagery.135 While in many cases technologies are still undergoing testing, industry use-
cases are numerous. Large distributors such as Amazon and Ocado have introduced
robotized and almost fully automated warehouse operations systems, where robots
stack shelves and autonomously prepare delivery orders.136 Amazon has progressed
a step further, whereby it has begun publicly testing unmanned aerial vehicles
to deliver products to customers in the US and the UK.137 The technologies have
widespread applications within the financial sector, where it is playing an increasingly
prominent role in identifying and executing trades and granting loan agreements. In
tackling financial crime, AI is being deployed to prevent fraud, tax evasion and money
laundering through the detection of anomalies, such as suspicious purchases and
uncharacteristic transactions.138

On the demand-side, AI is being used to deliver personalization of goods and

services. Analysis and understanding of Big Data is enabling companies to better
understand their customers’ preferences and allocate their products to market
segments accordingly. Through the combination of social media, mobile applications
and e-commerce, consumer goods companies are better able to develop and distribute
their products.139 The use and purchase data from the services of these companies
is being used to reinforce the companies’ understanding of their customers and as
a result, improve the targeting of their goods and services to the respective market
segments. As a result, consumers are receiving more relevant products and services.

135 Andrew Scott et al., “Modeling Artificial Intelligence and Exploring Its Impact” (Frederick S. Pardee Center for
International Futures Josef Korbel School of International Studies University of Denver, May 2017),
https://pardee.du.edu/sites/default/files/ArtificialIntelligenceIntegratedPaper_V6_clean.pdf.
136 James Vincent, “Welcome to the Automated Warehouse of the Future,” The Verge, May 8, 2018,
https://www.theverge.com/2018/5/8/17331250/automated-warehouses-jobs-ocado-andover-amazon.
137 April Glaser, “Why Amazon Is Testing Drone Delivery in the U.K. — and Not in the U.S.,” Vox, December 14,
2016, https://www.vox.com/2016/12/14/13955818/amazon-drone-delivery-uk-us-faa-testing.
138 E. Geist and A. J. Lohn, How Might Artificial Intelligence Affect the Risk of Nuclear War? (rand.org, 2018),
https://www.rand.org/pubs/perspectives/PE296.html.
139 Bernard Marr, “27 Incredible Examples Of AI And Machine Learning In Practice,” Forbes, accessed July 15, 2019,
https://www.forbes.com/sites/bernardmarr/2018/04/30/27-incredible-examples-of-ai-and-machine-learning-in-
practice/.

60 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
As a result of the improvements in supply chains and better market understanding,
AI has the potential to be a crucial economic stimulant. The overall impact of
automation and AI-related activities is not evident, but the benefits of AI are already
visible in the energy, transport, finance, manufacturing and retail sectors. On the
supply side, implementation of AI is leading to the automation of processes and
augmentation of worker experience, in turn generating efficiency gains in production.
On the consumer side, targeted advertising is processing mass amounts of customer
data and applying algorithms to improve the classification of its audience for targeted
advertising campaigns. Resulting gains stimulate innovation and drive international
trade due to increased competition both in hardware, software, services and consumer
goods. AI technologies should further accelerate the transition to services economies
and boost output of products such as robots, microchips, sensors and other machine
equipment.140 Furthermore, economic growth will result from required upgrades in
current (telecommunications) infrastructure, as well as the addition of internet fiber-
optic cables and the roll out of 5G network equipment around the world.141

The dual-use nature of AI naturally presents economic opportunities, which are

however, evidently outweighed by the geopolitical implications of the economic risks
presented by this range of technologies. Many current projections are hypothetical
estimates as the technology is evolving exponentially, thus immediate conclusions
and the extent of the impact cannot be determined with certainty.142 As a result, the
threats of AI seem more apparent than the opportunities. While the benefits currently
appear as narrow process-augmenting functions, they have the potential to have
immense overall impact on the economy.

2.2 Sociopolitical
As AI technology reaches into every aspect of daily routines, it is set to augment
the basic principles of societal functions. Commercial interests have generated
a path dependence in technological progress, as companies seek to understand
the most intricate details about their customers. As a result of their dominance
in the development of AI, intricate algorithms and data generating tools are now
being applied en-masse, in use cases which span far beyond targeted advertising.
AI is enabling governments to understand their citizens better, in some cases for
the improvement of the quality of life, while in others, for stemming individual

140 Joshua P. Meltzer, “The Impact of Artificial Intelligence on International Trade,” Brookings (blog), December 13,
2018, https://www.brookings.edu/research/the-impact-of-artificial-intelligence-on-international-trade/.
141 “Do We Understand the Impact of Artificial Intelligence on Employment? | Bruegel,” accessed July 15, 2019,
https://bruegel.org/2017/04/do-we-understand-the-impact-of-artificial-intelligence-on-employment/.
142 Steve Lohr, “A.I. Will Transform the Economy. But How Much, and How Soon?,” The New York Times, November
30, 2017, sec. Technology, https://www.nytimes.com/2017/11/30/technology/ai-will-transform-the-economy-
but-how-much-and-how-soon.html.

Macro Implications of Micro Transformations 61

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 freedoms. AI is enabling enhancements of online experience and augmentation
of social interactions. In states with limited protection of human rights, AI is set
to empower ruling governments to strengthen their control over their population
and project their influence abroad. This results in AI-enforced governance models
at home, and export of digital totalitarianism and diversified foreign influence
campaigns abroad, which reinforce existing social polarization. In these cases, AI is
tilting the tug-of-war between citizens and states in favor of the latter. A high-level
overview of the threats and opportunities associated with AI’s application within the
sociopolitical domain is provided in Figure 8 below.

Threats Opportunities
Centralization of data 
AI-enforced governance models   Improvements in the quality of life
Export of digital totalitarianism   Enhancements to the online experience
Reinforcement of societal polarization   Augments social interactions
Diversification of foreign influence campaigns 

Sociopolitical

Figure 8: Sociopolitical Threats and Opportunities of AI.

2.2.1 Threats

As highlighted in the economic section, the use of AI by leading data-generating

companies to reinforce their market position has socio-political connotations.
Centralization of data, whether by companies or governments, is establishing the
intricate understanding of customers or citizens behaviors’ by select entities.143 In
a socio-political context, this acts as a foundation for AI-driven surveillance and
control of citizens, as is becoming evident with China’s experimentation with digital

143 Jathan Sadowski, “Companies Are Making Money from Our Personal Data – but at What Cost? | Technology
| The Guardian,” The Guardian, 2016, https://www.theguardian.com/technology/2016/aug/31/personal-data-
corporate-use-google-amazon.

62 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
social credit systems.144 Meanwhile, in Western democracies, the dependence of
large swathes of the population on services provided by the FAANG corporations has
stalled anti-competition legislation against these countries both in North America and
Europe.145 In both liberal and illiberal states, the growing dependence on platforms
that centralize data collection increases the risk of infringement of human rights, by
means depicted in the following paragraphs.

The rise of AI-enforced governance models and the export of digital totalitarianism
are intimately linked, largely because the implementation of AI-enforced governance
models requires the development of easily transferable technologies.146 Within the
context of this study, these phenomena have been predominantly observed in the
Russian and Chinese case studies. In the Chinese case, the combination of AI-driven
facial recognition, centralization of communication platforms (WeChat, etc.) and
state utilization of ‘ground level’ (healthcare, etc.) data has allowed for the inception of
an early social credit system in which citizens are awarded credit for ‘good behavior’,
and penalized for ‘bad behavior’, including, among others, jaywalking and walking
a dog without a leash.147 Penalties reportedly range from citizens being precluded
from using certain forms of public transport, denying their children enrollment into
the best schools, and in the worst cases, to imprisonment.148 Russia utilizes a similar,
though far less-developed system to optimize its domestic information campaigns,
and to identify and repress dissent domestically.149 The technologies underlying these
control systems are easily exportable, and – in serving to consolidate the regimes
which implement them – have the potential of ‘normalizing’ illiberal governance
internationally.150 Armenia, Azerbaijan, Belarus, Ecuador, Kazakhstan, Kenya, Pakistan,
Singapore, Sudan, Tunisia, Venezuela and Zimbabwe have all sought to implement
Chinese surveillance technologies in bids to emulate Beijing’s model of AI-led
repressive governance.151

144 Jack Karsen and Darrel M. West, “China’s Social Credit System Spreads to More Daily Transactions,” Brookings
(blog), June 18, 2018, https://www.brookings.edu/blog/techtank/2018/06/18/chinas-social-credit-system-
spreads-to-more-daily-transactions/.
145 John Naughton, “Tech Giants Face No Contest When It Comes to Competition Law,” Then Guardian, 2017,
https://www.theguardian.com/commentisfree/2017/jun/25/tech-giants-no-contest-on-competition-law-
amazon-whole-foods.
146 Amy Hawkins, “Beijing’s Big Brother Tech Needs African Faces,” Foreign Policy (blog), accessed June 28, 2019,
https://foreignpolicy.com/2018/07/24/beijings-big-brother-tech-needs-african-faces/.
147 Alexandra Ma, “China has started ranking citizens with a creepy ‘social credit’ system — here’s what you can do
wrong, and the embarrassing, demeaning ways they can punish you,” Business Insider Nederland, October 30,
2018, https://www.businessinsider.com/china-social-credit-system-punishments-and-rewards-explained-2018-4.
148 Vicky Xiuzhong Xu and Bang Xiao, “‘Punishing the Disobedient’: China’s Social Credit System Could Engineer
Social Behaviour by 2020,” Text, ABC News, March 31, 2018, https://www.abc.net.au/news/2018-03-31/chinas-
social-credit-system-punishes-untrustworthy-citizens/9596204.
149 Soldatov and Borogan, “Russia’s Surveillance State.”
150 Wu’er Kaixi, “China’s New World Media Order | by Wu’er Kaixi & Christophe Deloire,” Project Syndicate, June 3,
2019, https://www.project-syndicate.org/commentary/china-press-freedom-attack-democracy-by-wu-er-kaixi-
and-christophe-deloire-2019-06.
151 For the full list of countries, see Freedom House, “Freedom on the Net 2018: The Rise of Digital Authoritarianism”
(Freedom House, October 31, 2018), https://freedomhouse.org/sites/default/files/FOTN_2018_Final%20
Booklet_11_1_2018.pdf.

Macro Implications of Micro Transformations 63

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 A linkage is also evident between the previously identified societal polarization in
liberal democracies and increases in the impact of diversified foreign influence
campaigns. Though autocrats take an active role in propagating ‘threatening’ AI-
related applications, several of the liberal democracies included within this study,
with the United States in particular, can be credited with facilitating the propagation
of socio-politically-relevant AI technologies. These derive largely from big tech
companies’ use of clustering models to maximize advertising revenues, and can
be universally associated with the exacerbation of political radicalization in liberal
democracies.152 This is because the clustering models are used not only to present
users with relevant advertisements, but also to maximize their time spent on the
platform by, among others, filtering the news articles they interact with, the ‘groups’
the platform recommends joining and the ‘influencers’ they are encouraged to follow.
Big tech companies, such as Twitter and Facebook, increasingly create focal points
within their users’ content consumption habits.153 These algorithms impact the
health of the discourse in liberal democracies as they cluster significant cohorts of
populations who are led to perceive the world in largely segregated media ecosystems.154
The creation of these so-called echo chambers undermines social cohesion and breeds
societal polarization by cementing identities.155 The EU has played an active role in
attempting to curtail the impact of this phenomenon, through, among others, the
implementation of the GDPR.

Another issue gaining prominence in the reinforcement of social polarization is

the rise in augmented video, image and audio content, known as deepfakes.156 While
liberal democracies have been identifying methods to counter written disinformation
on social media, development of deepfakes has matured and is easily accessible to
the average internet user. Deepfakes are developed using the generative adversarial
networks (GAN) AI technology, where algorithms are trained to replicate patterns
such as the face or voice of a political figure.157 Recent breakthroughs have made the
creation of deepfakes as simple as text editing, whereby an algorithm matches the text
to create a video or audio based on past video or audio samples.158 Deepfakes have in

152 Matt Kapko, “How Social Networks Are Changing Mobile Advertising,” CIO, August 20, 2014, https://www.cio.
com/article/2475406/how-social-networks-are-changing-mobile-advertising.html; Mostafa M. El-Bermawy,
“Your Filter Bubble Is Destroying Democracy,” Wired, November 18, 2016, https://www.wired.com/2016/11/filter-
bubble-destroying-democracy/.
153 Kapko, “How Social Networks Are Changing Mobile Advertising.”
154 “AI, China, Russia, and the Global Order: Technological, Political, Global, and Creative Perspectives.”
155 Jeff Stibel, “Fake News and Social Media: Confirmation Bias Puts Us in Echo Chambers,” UsaToday, May 15, 2018,
http://www.usatoday.com/story/money/columnist/2018/05/15/fake-news-social-media-confirmation-bias-echo-
chambers/533857002/.
156 Villasenor, “Artificial Intelligence and Bias.”
157 Karen Hao, “Inside the World of AI That Forges Beautiful Art and Terrifying Deepfakes,” MIT Technology
Review, accessed June 28, 2019, https://www.technologyreview.com/s/612501/inside-the-world-of-ai-that-forges-
beautiful-art-and-terrifying-deepfakes/.
158 Stanford University, “Edit Video by Editing Text,” Stanford News, June 5, 2019, https://news.stanford.
edu/2019/06/05/edit-video-editing-text/; Emerging Technology from the arXiv, “Facebook’s AI System Can
Speak with Bill Gates’s Voice,” MIT Technology Review, accessed June 28, 2019, https://www.technologyreview.
com/s/613647/facebooks-ai-system-can-speak-with-bill-gatess-voice/.

64 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
recent years been used in political contexts, in one instance by the Flemish Socialist
Party’s video campaign of a faked Donald Trump speech on climate change.159 Deep
fakes are therefore yet another tool to reinforce societal polarization and enable
disinformation campaigns to be more impactful, as visual content has a higher
tendency to go ‘viral’ on social media platforms and as a result, capable of reaching
wider audiences. Moreover, the ease of access of deepfakes on the internet and the
offerings of deepfakes-as-a-service presents them as a powerful tool for non-state
actors that are seeking to generate the greatest impact at the lowest possible cost.
As such, deepfakes are an example of AI shifting the balance of power, particularly
between state and non-state actors.

Collectively, AI technologies are bolstering the socio-political projections by illiberal

regimes, by propagating their rhetoric more effectively, both domestically and abroad.
China and its western Xinjiang province in particular have been testing grounds
for AI-enforced governance models. This has attracted interest from numerous
countries seeking to establish population control methods in fear of increased social
mobilization and popular revolutions. The likely result is the continued splintering of
the internet along governance models and further isolation of authoritarian regimes,
highlighted by China’s online censorship and Russia’s continued efforts to develop the
ability to disconnect itself from the global internet.160 Meanwhile, AI is increasingly
leveraged to fortify digital echo chambers and provide disinformation campaigns
with more tools to stifle debate on controversial issues in liberal democracies. The
complacency of Western big tech firms in both selling technology to authoritarian
regimes with limited due diligence and in removing disinformation content
from social media platforms has aggravated the issue further.161 Finally, as liberal
democracies seek to resist and/or counter disinformation and challenge authoritarian
narratives, they themselves risk weakening own principles in competition with less-
morally oriented counterparts. The erosion of basic human rights is therefore a risk
in liberal democracies too, where recent legislative action, such as the EU’s GDPR has
sought to constrain the capabilities of the big tech in collection of consumer data.

2.2.2 Opportunities

AI-driven functions are already contributing to various aspects of societal functions

and improvements in the quality of life. Demographic trends and projections across

159 Hans von der Burchard, “Belgian Socialist Party Circulates ‘Deep Fake’ Donald Trump Video,” POLITICO, May
21, 2018, https://www.politico.eu/article/spa-donald-trump-belgium-paris-climate-agreement-belgian-socialist-
party-circulates-deep-fake-trump-video/.
160 Charlotte Jee, “Russia Wants to Cut Itself off from the Global Internet. Here’s What That Really Means.,” MIT
Technology Review, accessed June 28, 2019, https://www.technologyreview.com/s/613138/russia-wants-to-cut-
itself-off-from-the-global-internet-heres-what-that-really-means/.
161 Dan Sabbagh, “Mark Zuckerberg Has ‘No Plans’ to Go to UK to Give Evidence to MPs,” The Guardian, May 15,
2018, sec. Technology, https://www.theguardian.com/technology/2018/may/15/mark-zuckerberg-facebook-
plans-uk-evidence-mps-parliament.

Macro Implications of Micro Transformations 65

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 the developed world are tending towards an increasing dependency ratio, where
the number of dependents supported by working individuals will continue to rise,
in part due to the rapidly ageing population in some states. Shortages of caretakers
and insufficient tax revenues in the future have prompted a search for alternatives,
with robotics being one of the pursuits.162 Japan has begun experimenting with
robots to replace human nurses, as well as wearable devices to support elderly
in care. In medicine, algorithms, are being trained to identify skin cancers from
imagery and have achieved detection rates comparable to dermatologists. Similarly,
algorithms developed to detect prostate cancer have already superseded a majority of
pathologists, with correct positive rates of 70% and 61%, respectively.163 As a result, AI
will increasingly be able to assist doctors in clinical decision support in complicated or
uncertain cases.

Beyond nursing and medicine, applications have been identified in legal processes,
law enforcement and education. Automation of legal tasks, such as the use of NLP to
process legal evidence, has been estimated to reduce the number of working hours
for lawyers by 2.5% annually over the next 5 years and as a result, expedite legal
proceedings.164 In law enforcement, adoption of computer vision has enabled the use
of facial recognition for the identification of suspects or wanted criminals, as well as
for improving border security.165 Meanwhile in education, NLP and machine learning
algorithms are being developed to personalize teaching in schools and expand class
sizes in higher education institutions.166 With the help of AI, schools have the ability to
introduce concepts such as ‘adaptive learning’ that help stimulate students and boost
their academic performance, producing, as a result, more agile future generations.
These technological advancements are enabling efficiency gains in delivery, expanding
provision and improving the quality of social services.

Beside opportunities for basic societal functions, AI plays a role in enhancing

people’s online experiences and augmenting social interactions, by, for instance,
prioritizing advertisements based on their online behavior, preferences and social
(inter-)actions. Consequently, this results in the appearance of more relevant products
to the individual. Another advantage offered by this process of personalization

162 Mizuho Aoki, “Nursing Care Workers Hard to Find but in Demand in Aging Japan,” The Japan Times Online, June
27, 2016, https://www.japantimes.co.jp/news/2016/06/27/reference/nursing-care-workers-hard-to-find-but-in-
demand-in-aging-japan/.
163 Yoav Shoham et al., “Artificial Intelligence Index: 2018 Annual Report” (Stanford, CA: AI Index Steering
Committee, Human-Centred AI Initiative, Stanford University, December 2018).
164 Jason Tashea, “Courts Are Using AI to Sentence Criminals. That Must Stop Now,” Wired, April 17, 2017,
https://www.wired.com/2017/04/courts-using-ai-sentence-criminals-must-stop-now/; Steve Lohr, “A.I. Is Doing
Legal Work. But It Won’t Replace Lawyers, Yet.,” The New York Times, March 19, 2017, sec. Technology,
https://www.nytimes.com/2017/03/19/technology/lawyers-artificial-intelligence.html.
165 Standing Committee on the One Hundred Year Study of Artificial Intelligence, “Artificial Intelligence and
Life in 2030” (Stanford, CA: Stanford University, 2016), https://ai100.stanford.edu/sites/g/files/sbiybj9861/f/
ai100report10032016fnl_singles.pdf.
166 Standing Committee on the One Hundred Year Study of Artificial Intelligence.

66 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
is that it enables users to connect with various other individuals that share the
same ideals, interests or lifestyles, and access topics relevant to their interests. The
ability to discover like-minded individuals and organize into groups using social
media platforms fosters more collective action that can appear in the form of social
movements based on certain beliefs or issues, as illustrated in movements such as
Fridays For Future. As in other contexts, however, AI is a dual-use technology, meaning
that while some algorithms drive collective action on pressing challenges such as
climate change, others are fortifying digital echo-chambers of extreme political groups
by directing individuals to more extreme content.

AI technologies are poised to transform the most basic social processes, from legal
proceedings and law enforcement to medical care and education. While most are
currently undergoing testing, the demonstrated potential is extensive. Machines are
increasingly outperforming human counterparts in linear, but complex tasks, allowing
the outsourcing of highly-specialized activities to AI and resulting in the improvement
of service delivery. Meanwhile, internet users are set to experience enhanced digital
environments, as collected user information is fueling classification algorithms
to provide content they are likely more interested in. In turn, this opens wider
opportunities for association and supports social movements.

2.3 Military-security
The military domain is often the early-adopter of nascent technologies, fueled by
extensive funding and the pursuit to outperform adversaries. As such, the use-cases
of AI technologies in the military are clearly identified, as the deployment of such
technology possesses the ability to reduce military as well as civilian casualties
and alter warfare as a whole. Moreover, AI presents an opportunity in reduction of
operational costs and operational efficiency gains. On the other hand, the risks it
presents are reflected in both combat settings and at the geopolitical level, where an
arms race between leading AI-developing countries is gaining steam. The increasing
permeation of AI in the military domain is leading to fewer constraints on going
to war, with friction breeding escalation. The asymmetric nature of AI application
is leading to upsets in the military balance of power, resulting in the possibility of
hyper war with the ‘human-out-of-the-loop’. A high-level overview of the threats
and opportunities associated with AI’s application within the military-security domain
is provided in Figure 9 below.

Macro Implications of Micro Transformations 67

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 Threats Opportunities
Upsets military balance of power 
 Preservation of life
Friction breeds escalation 
 Reduced operating costs
Hyper war and human out of the loop 
 Increases operational efficiency
Fewer constraints on going to war 

Military-security

Figure 9: Military-security Threats and Opportunities of AI.

2.3.1 Threats

There are a number of threats associated with the inclusion of AI in existing military
capability portfolios. The first is that new AI powered systems can upset the existing
military balance of power by making traditional systems and doctrines obsolete.
Development of AI has presented an alternative stream of competition, providing
militarily stagnant countries with a new opportunity to re-enter the race. This has
been the case with Russia and China, which, despite spending on military a fraction
of what the US’ does, have concentrated large proportions of the spending in AI
research.167 China’s accelerated rate of AI development has positioned it to outpace
the US, and considering the close proximity of Chinese enterprises with the Chinese
state, the technological gains are highly likely to be transferred to the military.
Moreover, there is a diffusion of power to non-state actors, that are able to utilize
low-cost, widely available tools to sow discord. This can be carried out, among other

167 Christina LarsonFeb. 8, 2018, and 9:00 Am, “China’s Massive Investment in Artificial Intelligence Has an
Insidious Downside,” Science | AAAS, February 7, 2018, https://www.sciencemag.org/news/2018/02/china-s-
massive-investment-artificial-intelligence-has-insidious-downside; Gregory Allen Kania Elsa B., “China Is Using
America’s Own Plan to Dominate the Future of Artificial Intelligence,” Foreign Policy (blog), accessed June 28,
2019, https://foreignpolicy.com/2017/09/08/china-is-using-americas-own-plan-to-dominate-the-future-of-
artificial-intelligence/; Debby Wu, Henry Hoenig, and Hannah Dormido, “Who’s Winning the Tech Cold War?
A China vs. U.S. Scoreboard,” accessed June 28, 2019, https://www.bloomberg.com/graphics/2019-us-china-who-
is-winning-the-tech-war/.

68 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
ways, through the use of deepfakes in disinformation campaigns or the use of cheap
unmanned drones to provoke or battle more sophisticated adversaries.168

Among the major contestants for AI-driven systems, friction breeds escalation. The
US, Chinese and Russian governments’ focus on exploiting AI-related technologies
within the military highlights the security dilemma associated with not exploring
these weapons.169 The US and Chinese approaches differ markedly, with the Chinese
military focusing more heavily on empowering algorithms to partake in strategic
decision making, and the US military being geared almost solely to developing
‘human-in-the-loop’ systems. The asymmetry between these approaches incentivizes
continued investment, as both strive to overcompensate for the activities of the other.170
The upsetting of the military balance of power is associated with increased friction
and escalation potential within the military-security domain, as countries like Russia
and China attempt to challenge the status quo, while the US retrenches to maintain
its military superiority.171

The large-scale advent of an AI arms race significantly increases the risk of a hyperwar
with the ‘human-out-of-the-loop’, as the pressures associated with ‘winning’ such
an arms race may incentivize states to cut corners in pre-deployment testing and
potentially declaring the technologies ‘operational’ prematurely. AI technology will
expedite decision-making processes and make split decisions, resulting in narrower
time frames for non-military crisis management in times of conflict escalation.
Because autonomous (‘human-out-of-the-loop’) weaponry can operate independent
of human oversight,172 these systems can also escalate the pace of conflict operations
to such a degree that human intelligence can no longer keep up, resulting in the
advent of so-called ‘hyperwars’.173

The role of AI in combat is profound, as the political costs of going to war decrease
with the prospect of reduced human costs. Deployment of unmanned equipment
reduces the loss of soldiers in combat, hence resulting in fewer constraints on going
to war. The 2015 shootdown of a Russian fighter jet by the Turkish Air Force near the
Turkey-Syria border resulted in a major diplomatic fall out between the two nations,
including responses at the head-of-state level. Meanwhile, multiple takedowns of

168 “Human-Machine Teaming” (Ministry of Defence, United Kingdom, 2018), https://assets.publishing.service.

gov.uk/government/uploads/system/uploads/attachment_data/file/709359/20180517-concepts_uk_human_
machine_teaming_jcn_1_18.pdf.
169 This is because AI weapons face states with a modern-day ‘security dilemma’. See Matthijs M. Maas, “How Viable
Is International Arms Control for Military Artificial Intelligence? Three Lessons from Nuclear Weapons,”
Contemporary Security Policy, February 6, 2019, 1–27, https://doi.org/10/gfz53m.
170 Pecotic, “Whoever Predicts the Future Will Win the AI Arms Race.”
171 Larry Lewis and Anna Williams, “Impact Of Unmanned Systems To Escalation Dynamics” (CNA, n.d.),
https://www.cna.org/CNA_files/PDF/Summary-Impact-of-Unmanned-Systems-to-Escalation-Dynamics.pdf.
172 Boulanin and Verbruggem, “Mapping the Development of Autonomy in Weapon Systems.”
173 John Allen et al., “Future War NATO? From Hybrid War to Hyper War via Cyber War,” GLOBSEC NATO Adaptation
Initiative (GLOBSEC), accessed April 14, 2019, https://www.globsec.org/wp-content/uploads/2017/10/GNAI-Future-
War-NATO-JLF-et-al.pdf.

Macro Implications of Micro Transformations 69

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 Iranian drones in Syrian airspace by American forces since 2017 resulted in no more
than a heated exchange between Russian and US military officials.174 As a result,
engagement against unmanned equipment reduces the threshold for adversaries to
engage in combat. Moreover, in terms of upsetting the military balance of power,
non-state actors are able to counter state actors more effectively. In 2017, the US
army reported a US ally using a $3 million Patriot missile system to disable an enemy
quadcopter, which happened to be an inexpensive commercially available drone.175
This instance demonstrates the changing nature of combat with the introduction of
AI-powered equipment, which empowers non-state actors and challenges traditional
weapons systems and tactics. The issue is set to gain further prominence, as AI enables
the development of ‘swarm’ combat tactics, whereby unmanned systems will be able
to communicate with each other in combat.176

In summary, AI in the military domain, particularly when coupled with robotic

systems, has the potential to unsettle the current military balance of power by
enabling weaker state and non-state actors to challenge the status quo. The reducing
role of humans ‘in the loop’ decreases the threshold of attack and breeds friction,
as countries are more likely to test each other with no human cost involved. The
expedited decision-making processes in instances of military friction risk further
exacerbating the situation and leading to hyper wars. As a result, new or updated
principles and arms control regimes will be necessary to constrain the military
applications of the exponentially developing field of AI technologies.

2.3.2 Opportunities

Enhancing the safety of troops and the reduction of casualties in combat, both
military and civilian, is considered to be one of the primary objectives for the adoption
of AI in the military.177 In fulfilling operational tasks, AI-related technologies also
have the potential of directly contributing to reduced combatant attrition rates. This
is presented by these technologies’ ability to operate autonomously in hazardous
and inaccessible areas, as well as carry out terrain analysis and automated target
acquisition.178 In high-risk environments, autonomous weapons systems can be
deployed instead of soldiers, to reduce the risk of loss of life within own forces.

174 Michael R. Gordon, “American Warplane Shoots Down Iranian-Made Drone Over Syria,” The New York Times,
June 20, 2017, sec. World, https://www.nytimes.com/2017/06/20/world/middleeast/american-warplane-shoots-
down-iranian-made-drone-over-syria.html.
175 “Human-Machine Teaming.”
176 “Human-Machine Teaming.”
177 For an overview of many non-kinetic applications of AI, see De Spiegeleire, Maas, and Sweijs, “Artificial
Intelligence and the Future of Defense: Strategic Implications for Small and Medium-Sized Force Providers.”
178 Michael C. Horowitz, “Artificial Intelligence, International Competition, and the Balance of Power,” Texas
National Security Review 1, no. 3 (May 15, 2018), https://doi.org/10.15781/T2639KP49; “Daksh Remotely Operated
Vehicle (ROV),” Army Technology (blog), accessed July 15, 2019, https://www.army-technology.com/projects/
remotely-operated-vehicle-rov-daksh/; U. S. Army Training and Doctrine Command and Army Capabilities
Integration Center, The U.S. Army Robotic and Autonomous Systems Strategy, n.d.

70 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
The introduction of AI technologies, such as satellite image analysis, can be used to
identify civilians and diminish accidental killing of non-combatants, particularly in
air campaigns. Similarly, computer vision allows for the extraction and identification
of ‘objects of interest’ from moving or still imagery. This is illustrated in the US
Department of Defense’s Project Maven, which has the ability to improve targeting of
drone strikes, identify civilians and diminish accidental killing of innocent people.179

AI has the potential to be a transformative technology in generating operational

efficiency gains in the military. AI applications are associated with a range of
opportunities within tasks comprising intelligence, planning and mission support, as
well as combat operations. AI can improve the quality and speed of military planning,
analysis, forecasting and decision-making through the adoption of new technologies.180
As a result, AI offers the prospect of the preservation of life as well as reduction in
costs and increase in efficiency of operations. AI-enhanced functionality is especially
useful in critical situations where human responses are cognitively insufficient, or
when repetitive tasks – such as the monitoring of sensors – can be automated to
allow humans to dedicate their time to higher level tasks.181 Narrow application AI
systems have the potential to oversee combat operations through the processing of
large datasets to predict human action.182 The speed at which AI systems can process
information is particularly vital in the military-security domain as time-sensitive
decisions present a critical challenge to successful military operations.183 Furthermore,
the deployment of AI technology in a military context can fundamentally change the
character of warfare, resulting in a transformation from what the Chinese military
has identified to be today’s “informatized” ways of warfare to future “intelligized”
warfare.184 This will accelerate the cognitive speed in decision-making and thus
improve situational awareness through means such as taking humans “out of the loop”
in robotic and autonomous systems.

179 Cheryl Pellerin, “Project Maven to Deploy Computer Algorithms to War Zone by Year’s End,” U.S. Department
of Defense, July 21, 2017, sec. DoD News, Defense Media Activity, https://dod.defense.gov/News/Article/
Article/1254719/project-maven-to-deploy-computer-algorithms-to-war-zone-by-years-end/.
180 Horowitz, “Artificial Intelligence, International Competition, and the Balance of Power.”CEOs, and academics
have suggested that a revolution in artificial intelligence is upon us. Are they right, and what will advances in
artificial intelligence mean for international competition and the balance of power? This article evaluates how
developments in artificial intelligence (AI
181 Such as a swarm attack of unmanned devices or an inbound hypersonic weapon. See Andrew Feickert et al.,
“U.S. Ground Forces Robotics and Autonomous Systems (RAS) and Artificial Intelligence (AI): Considerations
for Congress,” Informative report (Congressional Research Service, November 20, 2018), https://fas.org/sgp/crs/
weapons/R45392.pdf.
182 Horowitz, “Artificial Intelligence, International Competition, and the Balance of Power.”CEOs, and academics
have suggested that a revolution in artificial intelligence is upon us. Are they right, and what will advances in
artificial intelligence mean for international competition and the balance of power? This article evaluates how
developments in artificial intelligence (AI
183 Feickert et al., “U.S. Ground Forces Robotics and Autonomous Systems (RAS) and Artificial Intelligence (AI):
Considerations for Congress.”
184 Elsa B. Kania, “数字化 – 网络化 – 智能化: China’s Quest for an AI Revolution in Warfare,” The Strategy Bridge,
accessed June 28, 2019, https://thestrategybridge.org/the-bridge/2017/6/8/-chinas-quest-for-an-ai-revolution-in-
warfare.

Macro Implications of Micro Transformations 71

This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms
 The use of AI in the military domain also proves to be economical, as it permits
the reduction of costs of defense and conducting military operations as well as
overall gains in the efficiency of combat tasks, such as equipment operation, target
acquisition and electronic warfare. For instance, deployed defensive systems, similar
to the Russian S-400 system and Israel’s Iron Dome, are able to analyze data about
the likely impact zone, use radar to track incoming short-range projectiles and fire
interceptor missiles. In this fashion, the system stops incoming missiles or aircraft
instantly, significantly faster than a human could react. Such systems not only
contribute to more efficient military operations, but also reduce the costs thereof. An
example is provided in swarms of autonomous, long-range, and low-cost kamikaze
drones, which upon becoming available, are expected to reduce the relevance of
aircraft carriers in conflicts of the future.185 Moreover, cost reductions in air warfare
will be driven by decreased cost of procurement of aircraft and piloting staff. In the
US, the training of an aircraft pilot is estimated to cost $557,000, while training a
drone pilot costs a mere $65,000.186 The same applies to actual aircraft, although with
unit costs varying depending on modification. The unit cost of a US MQ-9 Reaper
unmanned aerial drone is $6.5 million, while the unit cost of the latest generation F-35
Strike Fighter is between $148-$337 million (depending on the modifications).187 Given
these elements and the changing nature of combat, since 2014, the US Air Force is
training more drone pilots than fighter and bomber pilots combined.188 Therefore, the
cost saving opportunities presented by AI-driven systems are extensive, particularly
when incorporating the cost of losing soldiers in battle or the cost of long term
treatment for those injured in combat.

In conclusion, AI is transforming the dynamic of conflicts, by reducing both the

political and material cost of combat, while enhancing operational abilities of military
forces. This combination enables militaries to achieve their objectives at lower costs
and in new formats, such as through human-machine teaming. As a result of this, AI is
contributing to improvements the safety of soldiers and reducing civilian casualties in
high-intensity situations.

185 Greg Allen and Taniel Chan, “Artificial Intelligence and National Security,” National Security, 2017, 132.
186 “Dilbert at War,” The Economist, June 23, 2014, https://www.economist.com/united-states/2014/06/23/dilbert-at-
war.
187 Wayne McLean, “Drones Are Cheap, Soldiers Are Not: A Cost-Benefit Analysis of War,” The Conversation,
accessed June 28, 2019, http://theconversation.com/drones-are-cheap-soldiers-are-not-a-cost-benefit-analysis-
of-war-27924; War Is Boring, “How Much Does an F-35 Actually Cost?,” War Is Boring (blog), July 27, 2014,
https://medium.com/war-is-boring/how-much-does-an-f-35-actually-cost-21f95d239398.
188 “Dilbert at War.”

72 HCSS Report
This content downloaded from 41.141.59.162 on Wed, 20 May 2020 17:48:11 UTC
All use subject to https://about.jstor.org/terms