AI & WARFARE IN INDIAN CONTEXT : STATUS & RECOMMENDATIONS

“In the future high-end fight, we envision a world of algorithmic warfare and autonomy where competitive advantage goes to the side that understands how to harness 5G, AI, enterprise cloud, and quantum, when quantum is available, into a viable operational model”

Lt Gen Jack Shanahan, Director, Joint Artificial Intelligence Centre, US Military[1]

GENERAL

Warfare has been constant companion of mankind for thousands of years of both recorded and unrecorded history. On observing wars through history, all victors seem to have achieved better of opponent in any/ all of three important Strands of Superiority. These are Human Intelligence, Domain Superiority and Technology.

Among the trio, ‘Technology’ has a potential to change the very nature of warfare. Many of the new technologies such as Artificial Intelligence (AI), Quantum Technology, Autonomous Systems and 5G are inter-related with each other. In this article we focus on AI in particular.

PREVIEW

The article will cover the topic in following parts: -

• Who wins & why ?
• New Technologies.
• AI and Robotics: Global and Indian Perspectives.
• A Suggested Roadmap for India.

PART I: WHO WINS & WHY ?

Pursuance of Three Strands i.e. Human Intelligence, Domain Superiority and Technology in warfare to achieve victory, has been a running theme throughout history of mankind. The same is explained below:-

• Human Intelligence. From Big Bang[2] (13.7 billion years ago), origin of Homo Sapiens (2 lakh years ago), Agrarian Era (10,000 years ago) to present day, the pace of evolution of mankind has an important lesson. The complexity and scale of achievements of mankind have grown exponentially in last few thousand years. The primary reason for this quickened pace of growth were intelligence and mental capacity of humans. Since advent of computers, in last 80 years alone US and USSR were able to land on moon, conduct moon landings and even discover nuclear bomb thereby achieving great power status. In military field, the human intelligence manifests in the form of brilliant planning and execution by renowned commanders. In the age of computers, we have seen exponential results when intelligence is supported by satellite images/ drones/ encryption-decryption devices and such like things which essentially use computing. If computing support to human intelligence can make such a difference, what if intelligence itself is multiplied before being aided by computing? The list of possibilities will be endless. This 'intelligent computers' are essentially AI powered machines which leverage human intelligence as well as technology.
• Domains Superiority. The prevalent domains include Land, Sea, Air while modern domains are Cyber, Space, Cognitive and Electro-Magnetic Spectrum as visualised by various theorists of warfare. Historically land, followed by water were two contested domains till arrival of air force. Warring opponents in future are likely to aim for superiority in new Domain Operations for deterrence/ coercion in peace and victory in war.
• Technology. The third Strand of Superiority is technology. The opponent with disruptive technology wins. Historically, whether it is advent of Gun Powder, Rifles, Machine Guns, Tanks or Aircrafts, the superior technology has decisively impacted overall operations. Whether technology alone is sufficient, is another matter.
• Emerging Dynamics of Warfare. In future, technologies such as AI will enable expansion of Domains to include Cognitive, Cyber, Space and Electro Magnetic Spectrum. It will also transform existing domains with Autonomous Systems and processes. Tempo of Operations are likely to transform in a way that Human Intelligence may not be able to keep pace without Decision Support from AI based Quantum Computers.

Having seen the domains of superiority, we will focus on New Technologies in general and AI in particular as a strand of superiority that needs to be pursued by Indian Military.

PART II: NEW TECHNOLOGIES

Global Trends in New Technologies. The early Chinese interest in New Technologies resulted in China overtaking US[3] in key areas of AI and 5G mobile telephony. In fact PLA anticipates that AI may fundamentally change the very nature of warfare and transform it from “Informatised” to “Intelligentized” warfare.[4]. The head start spurred US to launch “Third Offset Strategy” to build capabilities to offset Chinese strength in these new technologies[5]. NATO Science and Technology Organisation’s Science and Technology Trends: 2020- 40[6] published in March 2020, lists eight emerging technologies with disruptive potential in next 20 years. The technologies are given in Figure 1 below: -

Figure 1: Emerging and Disruptive Technologies 2020–40

Inter-relations of New Technologies

New technologies are much inter-related. For instance, AI provides intelligence to Systems such as Autonomous Drone Swarms with Mission Command type of orders, the swarms can transfer battlefield surveillance/ damage assessment data through 5G clouds to Command Post (CP). Quantum Computers at CP can process data from multiple sources, this data is then analysed AI driven algorithmic process on Quantum Computers, which provide updated Decision Support or Automated Decision as per laid down policy. While much of this cycle is a visualisation (or rather work in progress), theoretically all of it is possible. The interplay of factors is given in Figure II below: -

Figure 2: Visualisation of Interplay of AI, Autonomous Systems/ Processes, Data Transfer and Quantum Computers

Before moving further, lets take a minute to understand Artificial Intelligence and how it has come to prominence.

• AI Defined. Professor John McCarthy of Stanford University considered as the founding father of AI, termed AI “as is the science and engineering of making intelligent machines, especially intelligent computer programs” and the “ultimate effort is to make computer programs that can solve problems and achieve goals in the world as well as humans”[7].
• The Game. AI received worldwide attention and praise after a game of Go in 2016. AlphaGo, a computer AI program developed by Google DeepMind, defeated Je Sedol, in a game of Go[8]. Je Sedol was 18 times World Champion of this game, when the matches were conducted. Go being an ancient Chinese board game of strategizing, the defeat had a major impact on Chinese psyche.
• Functions of AI. The functions of possible AI machines can be anything conceivable. Self-driving Cars, Household and Industrial Robots, Logistics Management System, Deep Space Exploration modelling and 3D printing are some of the widely known functions. Important military applications include Autonomous Systems and Processes, Decision Support Systems and Logistics Management Systems.

PART III: AI — GLOBAL AND INDIAN PERSPECTIVES

Global Perspective

• [10]. AI as an area of interest has come a long way. In 2019, nine percent of all published conference papers were related to AI, China being the world leader with 28 percent of global output[11]. Interestingly Chinese government institutions publish nearly three times more AI papers than the private Companies[12]. India ranks third in number of published papers pertaining to AI[13]. Global private investments in AI have risen to over $ 70 Billion and more than half of the amount is invested in Start Ups[14]. Here again, China is the world leader while India ranks dismal 10th[15].
• Technological Progress Areas in AI. Globally image recognition, speech recognition, language translation and data processing are the key areas where sizeable progress has been made. Image recognition helps in making Unmanned autonomous systems (such as Autonomous Cars and Drones), photo identification tools (being deployed aggressively by China in Xinxiang) and remote sensing (for change detection in a given area). Robot process automation was the fastest growing area in AI in 2018, with China as the world leader with more than 36 percent of all Industrial robots installed in China[16].

Military Interest Areas in AI

Intelligent Drones

• Intelligent Drones remain a primary focus area for Major militaries across the world. On 5 January 2018, two Russian military bases in Syria (including an Air Base and Logistic Base) were attacked by a swarm of 13 drones armed with explosives. Reportedly, seven drones were blown up with anti-aircraft missile systems, while the other six were hacked by defenders[17].
• In January 2019, a military parade in Saudi Arabia was targeted by Yemeni drone laden with explosives. It was successful in killing more than 25 personnel of Arab Coalition but more important was the psychological damage to sense of security of Coalition Forces[18].
• Drones with number of autonomous capabilities are already existing such as Israel’s Harpy loitering drone[19] and Chinese TYW-1 and ASN-216 drones with automated takeoff and landing capability. Swarms of up to 1000 drones have also been successfully tested in China[20].
• Intelligent Drones of future can be tasked for a number of functions including Drone Swarms, Loitering Munitions, Suicide Drones, Unmanned Fixed and Rotary Wing Aircrafts, Load Carrying drones used for dropping explosives and logistics resupply.

Intelligent Robots

• One of the prime concerns of Armies across the world is to enhance mobility of its troops particularly in obstacle ridden and mountainous terrain. Robots are being seen as an answer to this puzzle. This tasking adds to existing thoughts on application of Robots to be employed for intervention tasks in dangerous environments wherein a lot of systems are already under deployment. US army is developing a Robotic Mules with capability to carry up to 1000 pounds load for 60 miles range[21]. In addition, US Army is pursuing a separate program for Exo-skeletons as a measure to enhance combat efficiency of its troops.
• Chinese military is also developing separate Robots for re-supply and reconnaissance[22]. Future interest areas of Chinese Army in Robotics include participation of Robots in direct Combat in addition to support tasks already envisaged. Interestingly Chinese Army holds a biennial competition for land based unmanned systems since 2014. This competition scheduled for September 2020, has identified un-manned vehicle reconnaissance, materials loading and unloading, mobile power-supply, high-mobility transportation in mountain terrain, air-ground coordinated reconnaissance and strike of cluster unmanned systems[23] among other areas in which entries will be taken.

Decision Support Systems.

• This is a major area of research in AI. These systems are likely to be of great importance in reducing Cognitive Load of soldiers and Commanders in data rich environment of future, wherein a number of autonomous systems will provide mountains of data in real-time, which simply can’t be analysed and decided upon by a human being. US military’s Project Maven[24], which analyses video feed received from Unmanned Aerial Vehicles (UAVs). Its AI based algorithms are being trained to identify relevant data from these feeds and provide actionable inputs to the human operator. The program maybe a testbed for further integration of inputs from other sources to provide decision support.
• In China, AI based war gaming is a major area of research. In September 2017, in a military war game organised at National Defence University an AI based system known as CASIA-Prophet 1.0, defeated human teams by a score of 7 to 1.222 [25]. There are also indications that Joint Staff Department of Central Military Commission has asked PLA Commanders for using AI in Operational Planning and Decision Support[26].
• Realistic Training. Training is another area wherein militaries particularly PLA is using the AI in a big way. PLA is using Virtual and Augmented Reality for training of soldiers in simulated battlefield environments. It is also using creating an AI enabled Blue Force[27] for realistic wargaming.

Indian Perspective

• Governmental Initiatives. India as a nation has done reasonably well in the field of AI. Niti Aayog released a discussion paper on National Strategy for Artificial Intelligence[28] in June 2018.The paper primarily focuses on developmental aspects and identifies five sectors wherein AI tools may be deployed. These sectors include Agriculture, Education, Healthcare, Smart Cities and Infrastructure and Smart Cities and Mobility. In the paper there is no mention of defence at all.
• Private Sector[29]. India published third highest volume of AI Journal papers in AI in 2018 and it was among the countries with highest growth in AI hiring’s in 2019. It has the highest average penetration of AI skills globally with skills in all major sectors evenly distributed. Between July 2018 to July 2019, a total of 143 Start Ups from India received $ 360.1 Million in funding. In comparison 486 Chinese Start Ups received a total of $ 16.6 Billion in funding in the corresponding period which is roughly 46 times of Indian investment.

Developments in Defence Sector

• In February 2018, Ministry of Defence’s (MoD) constituted a task force under Tata Sons Chairman N. Chandrasekaran, to study the strategic implications of AI on National Security. The task force submitted its “Report of Task Force for AI” in June 2018[30]. The report is not available in public domain and further progress on it is not known.
• In January 2020, as per media reports, Defence Research Development Organisation (DRDO) is known to be establishing five dedicated labs for emerging technologies[31] such as AI, Quantum Technology, Cognitive Technology, Asymmetric Technologies and Smart Materials. Interestingly all these labs are to be led by young Scientists below 35 years of age, specially selected by a Committee led by Principal Scientific Advisor to the Prime Minister.

Having seen Global and Indian perspectives, some recommendations have been suggested in the next part for Indian Military to remain competitive in this emerging New Technological domain.

PART IV: RECOMMENDATIONS

The recommendations are based on principle of setting a direction followed by utilising existing strength in near term. In the near and medium term existing technologies and products should be test-bedded/ matured while long term plans should be to develop new technologies and products. The detailed recommendations are outlined in succeeding paras.

Short Term.

• A theoretical Concept for AI in Future Warfare should be laid down in India. In this regard an all-encompassing vision as put forward by Chinese PLA may be instructive rather than fragmented and rather adhoc Western ideas.
• Based on the Concept, an AI Vision for Armed Forces is required to be specified. It should include existing proven AI technologies and future possibilities. A priority-based timeline should be laid down for developing/ acquiring the same.
• A New Technologies Corpus of $1 Billion for supporting Research and Start Ups is strongly recommended. India has an excellent Start Up ecosystem already available. The amount may be a game changer in developing key products and technologies suited to military requirements.
• There is a need for establishment of a Tri Service Military Design Bureau which should have a separate cell of New Technologies for coordination with Academic Institutions, Research Centres and Industry. Based on Global scan and own vision, Bureau should identify Products and Processes for immediate operational requirements, these should be given to stakeholders as a Problem Statement.
• Drone Swarms, Robo-bots and land/ air based Robotic Mules should be test bedded taking advantage of easily available technology. They should be included in Operational Alerts for real life feedbacks.

Medium Term.

• Trials and induction of Drone Swarms, Robo-bots and land/ air based Robotic Mules should be commenced to further hone operational concepts and tactical employment.
• Taking advantage of progress in self-driving Cars in private industry, a Logistic Support System for military needs should be deployed. The logistic manpower should be reduced and budget saved should be utilised to support funding for research and product development in areas such as AI, Quantum Communication and 6G.
• India is likely to achieve reasonable progress in 5G mobile technology till then. A 5G cloud-based architecture with inbuilt redundancy should be planned for military communication and AI developmental requirements.
• Development of Autonomous Systems should be prioritised since 5G data architecture can support networking requirement of these systems. These include unmanned vehicles, aircrafts and submarines.
• Self-learning Decision Support Systems should be test-bedded with autonomous systems providing data for learning. A Red Force for war gaming should be developed.

Long Term.

• Autonomous systems and decision support systems should be deployed three services with human in the loop. Armed forces should be reduced substantially based on operational experience and funds diverted for further research and development of new technologies.
• Deploy Quantum technology along with 5G/ 6G data architecture for hackproof communication.
• Deploy space-based AI systems and Robots for surveillance and communication.

CONCLUSION

To borrow words from PLA’s Lieutenant General Liu Guozhi, director of the Central Military Commission’s Science and Technology Commission, “Facing disruptive technology, [we] must … seize the opportunity to change paradigms. Whoever doesn’t disrupt will be disrupted!”. AI and Robotics are two such disruptive technologies. To not be prepared is hardly an option.

[1] Big Data for Big Wars: JEDI vs. China & Russia, article by Sydney J Freeberg Jr dated 12 August 2019. Available at breakingdefence.com/2019/08/big-data-for-big-wars-jedi-vs-china-russia/ accessed on 17 May 2020

[2] Big History: Connecting Knowledge, online course by Prof David Christian and Mr David Baker of Macquire University, Australia. The online course was undertaken by author between 27 June to 04 August 2019 through Coursera mobile application.

[3] Ibid, refer page 6 for this extract on Chinese advances in AI overtaking US. “after Microsoft announced that its speech recognition software surpassed human level language recognition in October 2016, Andrew Ng, then head of research at Baidu, tweeted: “We had surpassed human-level Chinese recognition in 2015; happy to see Microsoft also get there for English less than a year later.” China also launched its New Generation AI Development Plan in July 2017 with stated aim to lead the world in AI by 2030.

[4] Battlefield Singularity: Artificial Intelligence, Military Revolution, and China’s Future Military Power by Elsa B. Kania, Published by Centre for New American Security in March 2017. Refer page 4 of document.

[5] Assessing the Third Offset Strategy, by Jesse Ellman, Lisa Samp and Gabriel Coll of Centre for Strategic and International Studies. See report (of discussion and analysis of subject held at CSIS on 28 October 2016.) for a good explainer of “Third Offset Strategy” and its genesis. See page 14 for understanding the Strategy itself.

[6] Science and Technology Trends: 2020- 40, published by NATO Science and Technology Organisation in March 2020. See pg vii of Executive Summary of the report for Emerging and Disruptive Technologies.

[7] Basic Questions (of AI) answered by Professor John McCarthy of Stanford University. He is considered the founding father of AI. The Professor, credited with visualizing AI and coining the term itself, expired in 2011. The questions and answers accessed at 2110 hrs on 18 May 2020, can be referred at this link http://jmc.stanford.edu/artificial-intelligence/what-is-ai/index.html

[8] Elsa B. Kania, citation number 3, PLA thinking on potential of AI, was said to be deeply influenced by results of these games (AlphaGo defeated Je Sedol in four games out of five. One game was won by Sedol). Since Go (WeiQi) is an ancient Chinese strategizing board game, an AI machine defeating a human was taken as signifying immense potential of machines in complex analysis and strategizing. Refer to page 15 of the document for references to PLA literature on the subject.

[9] Vincent Boulanin and Maaike Verbruggen, Citation number 19. Refer Page 91 of the Report. Report accessed on 19 May 2020. It can be downloaded from link https://www.sipri.org/sites/default/files/2017-11/siprireport_mapping_the_development_of_autonomy_in_weapon_systems_1117_1.pdf

[10] Artificial Intelligence Index: 2019 Annual Report, published by Stanford Human-Centered Artificial Intelligence, an independent initiative a Stanford University. The report is available at link https://hai.stanford.edu/research/ai-index-2019 . Report accessed on 20 May 2020. Data and statistics, pertaining to Global and Indian AI, have been liberally quoted from this report.

[11] Ibid, Refer page 2 and 15 of the report.

[12] Ibid, Refer page 16 of the report.

[13] Ibid, Refer page 16 of the report.

[14] Ibid, Refer page 24 of the report.

[15] Ibid, Refer page 89 of the report.

[16] Ibid, Refer page 104 of the report.

[17] Swarm Drone Attack in Syria Point to a New Kind of Warfare, by Mauro Lubrano on portal Global Risks Insights published on 18 January 2018. The article was accessed on 20 May 20, 2020. It is available on link https://globalriskinsights.com/2018/01/swarm-drone-attack-syria-uav/ .

[18] Houthi Drone Attack on Saudi Military Parade in Yemen Kills Top Coalition Officials, article published by portal NTHNEWS on 13 January 2019. Accessed on 20 May 2020. See accompanying video of Drone attack along with article available at https://nthnews.net/en/yemennews/houthi-drone-attack-on-saudi-military-parade-in-yemen-kills-top-coalition-officials/

[19] Mapping the Development of Autonomy in Weapon Systems, by Vincent Boulanin and Maaike Verbruggen, Report published in November 2017 by Stolkholm International Peace Research Institute (SIPRI). Refer page 53 of Report for capabilities of existing drones includinh Harpy.

[20] Elsa B. Kania, citation number 3, Refer page number 22 and 23 for a summary on PLA advances in Autonomous Drones and Swarm Drones.

[21] Soldiers soon to see robotic mules and tougher bomb bots in the field, by Todd South, published by portal Army Times on 22 November 2019. The article was accessed on 20 May 2020. It can be found on link https://www.armytimes.com/news/your-army/2019/11/22/soldiers-soon-to-see-robotic-mules-and-tougher-bomb-bots-in-the-field/ .

[22] Armor: Chinese Combat Droids Evolving, article published on portal StrategyPage on 17 May 2020. Accessed on 20 May 2020. Article available on link https://www.strategypage.com/htmw/htarm/20200517.aspx .

[23] PLA Army to host “Crossing Obstacles 2020” land-based unmanned system competition, published on China Military Online on 3 March 2020. Accessed on 20 May 2020. The article can be accessed on link http://eng.mod.gov.cn/news/2020-03/11/content_4861835.htm

[24] Palantir took over Project Maven Defense Contract after Google Backed Out, published by Portal AINews on 12 December 2019. The article was accessed on 21 May 2020. It is available on link https://artificialintelligence-news.com/2019/12/12/palantir-project-maven-defense-contract-google-out/ .

[25] Elsa B. Kania, citation number 3, Refer page number 28. PLA thinking on the subject envisages concepts like Battlefield Singularity, wherein the OODA loop will be so shortened by virtue of autonomous systems in multiple domains, quick data transfer (5G,6G and so on), data processing and decision support (Quantum Computing and AI) that human’s cognitive abilities simply will not be able to keep pace with speed of operations. Therefore a kind of singularity in battlefield will be achieved(see page 16 of this document for a better explanation of concept).

[26] Ibid, Refer page number 15.

[27] Ibid, Refer page number 28.

[28] National Strategy for Artificial Intelligence, discussion paper by Niti Aayog published in June 2018. The report was downloaded from website niti.gov.in on 20 May 2020. Focus areas have been covered between pages 24 to 45.

[29] Artificial Intelligence Index: 2019 Annual Report, citation number 9, Refer page 9 for data on AI Journal Papers, page 73 for hiring statistics, page 78–80 for skillset data and page 93 for data pertaining to Indian Start Ups.

[30] New Ministry of Defence Report Suggests Artificial Intelligence Can Provide “Military Superiority”, report edited by Sarthak Dogra was published on portal News18 Tech on 02 July 2018. It was accessed on 21 May 2020. The report is available on link https://www.news18.com/news/tech/new-ministry-of-defence-report-suggests-artificial-intelligence-can-provide-military-superiority-1797349.html

[31] DRDO to set up five new labs to prepare Indian armed forces for future high-tech warfare, article published on news portal Economic Times on 02 January 2020. Accessed on 21 May 2020. It is available on link https://economictimes.indiatimes.com/news/defence/new-drdo-labs-to-boost-indian-defence-capabilities/articleshow/73072702.cms