“AI5, ARTICLE 1, PRANJALI JOSHI, PROMISE AND PERILS OF ARTIFICIAL

INTELLIGENCE AND MACHINE LEARNING IN NATIONAL SECURITY”

“Artificial intelligence, machine learning, technology, defence, national security”

ABSTRACT

The rapidly growing artificial intelligence and machine learning are now making strides beyond
civil and commercial sectors. The technology is now heading towards military operations and is
changing the very nature of warfare, opening an arms race amongst the nations. Capable of
causing significant changes in our military capability, the technology can become a threat to
national security if given a cold shoulder. Hence, India must develop a robust to integrate AI and
ML in military operations so that the country remains prepared for the modern wars of
tomorrow.

In this Article Pranjali Joshi of New law college, Pune discusses the potential application of
artificial intelligence and machine learning in military operations, throws light on global
development, call intention to major impediments in the deployment and provides a way
forward.

ARTICLE

Artificial intelligence has penetrated every fathomable industry where the human mind is
engaged. It has revolutionized the standard business and individual model of labour and is now
being deployed beyond civil or commercial purposes. But the integration of artificial intelligence
and deep learning in military genomics has changed the very face of present-day wars. Many
developed nations are investing in AI based- research, a large part of which goes for defence
technology. This has given rise to an artificial intelligence arms race amongst the nations, mainly
the frontrunners- The US, Russia and China.

The relevance of using AI in defence gains importance in view of the spillover of the technology
to the non-state actors, examples being, drone strikes by Houthi rebels in Yemen and Ukrainian
Separatists in Ukraine respectively. The issue has become a concern for India as there have been
reports of Pakistan supported non-state actors using drones to cause instability in Indian regions
of Punjab, Jammu and Kashmir. Hence, it has become imperative for India to take a holistic
approach in developing the AI landscape, given the range of benefits it offers for a more
strengthened national security.

Application of AI in defence

Intelligence, surveillance and reconnaissance- AI is specifically useful for intelligence purposes

due to its efficient handling of large data sets and predictive analysis. iAI with geospatial systems
can help in identifying illegal or suspicious activity by extracting intelligence through footage
from radars and aerial vehicles.ii The extracted information then can be used to alert the
concerned authorities. In this capacity, AI can take over the work of human analysts, who
currently spend hours stifling through data to find suspicious information. Freeing analysts will
increase efficiency and reduce the time of decision making based on data. Some usages for this
purpose includes building algorithms for multilingual speech recognition, easy translation from
noisy environments, and fusing images to create 3-D models.

Logistics- AI may have great utility in the field of logistics. The predictive technology can be
employed to develop logistics information systems which can be used to check the condition of
vehicles, extent of damage and maintain a schedule as per the needs of individual vehicles.
iii
Considerable work has already been done using AI for logistics in the civilian sector, with
several Indian companies gaining expertise in this arena. Currently, the task of examining the
equipment is undertaken by technicians who inspect the vehicles and send them for repair.
Employing AI, will shorten the time period of repair which in turn will ensure time-efficient and
cost-effective logistics supply.

Cyberspace operations- Cyber Warfare as a modern means of weapon has become a challenge
for national security. A malicious network has the potential to severely undermine the stability
and security of the state. Artificial intelligence frameworks based on machine learning methods
can be used for the protection of critical information, namely defence applications and networks.
More specifically, hybrid evolving spiking anomaly detection models (HESADM) are systems
designed to prevent cyberattacks. These AI-enabled systems could detect, evaluate and patch
security vulnerabilities within seconds. This technology helps in preventing the system from
compromising, ensuring the security of the state as well as cyber operations. iv

Training- Training and computer simulation are cross-functional areas which may require
governments to develop models with the help of software engineering principles that will help
train soldiers in learning tactics which are used in actual military operations. The augmented and
virtual reality techniques can be particularly used for the purpose of creating realistic simulations
for training defence personnel. v

International use of AI in military operations

USA- The US Department of Defence released its first AI strategy in 2018 which aims to
increase use of AI in military operations from decision making to problem prediction through
investments. Some of the next generation AI projects deployed by the US military are- Squad X,
Project OFFSET, Low-Cost Unmanned Aerial Vehicle Swarm Technology (LOCUST) etc. The
most notable one being Project Maven- an independent command and control system to snoop on
targets. vi

China- Chinese government is also pressing for the infusion of AI and machine learning
techniques in military operations. China’s president President XI Jinping has emphasized on
pursuing a ‘military-civil fusion strategy' which has also been confirmed with the Chinese
Artificial Intelligence Development Plan. Some of their homegrown projects are- Military-Civil
Fusion for Artificial Intelligence Development programme under the aegis of Tsinghua
University, AI Baidu- has surpassed human levels of speech recognition, Blowfish A2 and UAV-
capable of performing automatic fixed-point timing detection and target precision strikes.vii

Military integration Challenges for India

Legal challenges- Military integration of AI can become a challenging task for the governments
from a legal perspective. LOAC- The law of armed conflict is the humanitarian law regulating
the humanitarian harm to the civilians. It is codified into treaties and protocols such as Four
Geneva Conventions, the Additional Protocols, and weapon specific treaties such as the Ottawa
Landmine Ban. Critics have argued that the artificial weapon systems are incapable of
differentiating between fearful citizens and hostile enemies. viiiHence, they may not be able to
understand and comply with the principles of humanitarian laws.

Ethical challenges- Use of AI poses ethical dilemmas such as accountability and human rights
violations. Accountability is a fundamental moral concept determining moral responsibility of
actions, which entails deciding reasonable retribution, emotions such as guilt and shame, and
making moral redressal.ix Therefore, accountability acts as an important deterrent, not just in war
but in social lives too. It has been contended that the autonomous weapon systems will take
decisions independently, thereby overlooking values, causing human rights violations. x

Operational challenges- There are several operational risks associated with the use of AI in the
military such as- accidents and data failure. The machines can at any time abruptly fail to
perform due to any technical issue. There are a myriad of examples of mechanical meltdowns
such as the Chernobyl disaster and the Space Shuttle Challenger disaster. Therefore, these
machines can be hardly trusted on the operational front. Moreover, these networks are vulnerable
to hacking, even the closed or air-gapped ones can be tampered for gaining access to confidential
information. xi

Conclusion

Undoubtedly, artificial intelligence and machine learning are the future of modern warfare. They
have the potential to revolutionize warfare and the military systems. But they can become a
challenge to national security if given a cold-shoulder. Therefore, India must embrace the
technology with open arms, because if we don’t our opponents will. The government should
meticulously work out the integration framework, keeping in mind the various challenges. The
focus should lay on investment by engaging in public-private partnership and international
cooperation. With wise integration, AI can strengthen the defence capability manifolds, reducing
labour costs and time of decision making. AI infusion in defense must be achieved so that the
country remains prepared for the battlefields of tomorrow.
i
USING ARTIFICIAL INTELLIGENCE IN BIG DATA, Analytics Insight, (2020),
https://www.analyticsinsight.net/using-artificial-intelligence-in-big-data/ (last visited Aug 4, 2021).

ii
Jean-Paul Yaacoub, Hassan Noura, Ola Salman,⁎ and Ali Chehab, Security analysis of drones systems: Attacks,
limitations, and recommendations, Volume 11, September 2020,
https://www.sciencedirect.com/science/article/abs/pii/S2542660519302112?via%3Dihub

iii
By Kevin Grigson, Group VP, Operations and Greg Boring, VP of Sales, How Predictive Analytics Is Transforming
Logistics Networks, https://blog.kencogroup.com/labor-shortage-impacts-mitigation-through-innovation-0

iv
Top 8 Machine Learning Tools For Cybersecurity, Analytics India, (2020) https://analyticsindiamag.com/top-8-machine-
learning-tools-for-cybersecurity/ (last visited Aug 4, 2021).

v
What Are The Scope and Challenges of Using AI in Military Operations, Analytics India, (2020),
https://analyticsindiamag.com/what-are-the-scope-and-challenges-of-using-ai-in-military-operations/

vi
Ibid.
vii
Ibid.
viii
Morgan, Forrest E., Benjamin Boudreaux, Andrew J. Lohn, Mark Ashby, Christian Curriden, Kelly Klima, and Derek
Grossman, Military Applications of Artificial Intelligence: Ethical Concerns in an Uncertain World, Santa Monica, Calif.:
RAND Corporation, RR-3139-1-AF, 2020. As of July 28, 2021: https://www.rand.org/pubs/research_reports/RR3139-
1.html

ix
Ibid.
x
Ibid.
xi
Musa G. Samaila, Miguel Neto, Diogo A. B. Fernandes, Mário M. Freire, Pedro R. M. Inácio, Challenges of securing
Internet of Things devices: A survey, issue 2, Volume 1, (March 2018),
https://onlinelibrary.wiley.com/doi/full/10.1002/spy2.20