Artificial Intelligence In Military Context
Artificial Intelligence, that includes Machine Learning, Deep Learning, etc is at the forefront of Industry 4.0. AI is gaining popularity in all domains to make the current systems and products more efficient and reliable, avoiding human intervention, hence eliminating human error. Article presents a deeper look into the world of AI and its applications in Defence.
Swarm technology by Botlab Dynamics
Technology superiority has been the major deciding factor in any military conflict since ages. Military capabilities have therefore always heavily relied on new technologies to avoid any surprise from their adversaries.
The industry 4.0 capabilities which are being developed and adopted by the industry, with a strong knowledge back up from the academia, are also showing their mark in the Military.
Amongst the different elements of Industry 4.0 (3D printing; Big Data; Autonomous Operations; Robotics - soft and hard; Nanotechnology; Augmented and Virtual Reality; Exoskeleton; Cloud Computing; Predictive Maintenance; Advanced Materials etc), Artificial Intelligence, that includes Machine Learning, Deep Learning, etc is at the forefront of Industry 4.0. AI is gaining popularity in all domains to make the current systems and products more efficient and reliable, avoiding human intervention, hence eliminating human error.
Commercial sector has recognized the immense potential of AI, with the large amount of funding it is receiving, vis-a-vis the military sector. There is a strong feeling that Military lags the commercial sector in this Knowledge Revolution. Taking this in a positive sense, militaries’ AI R&D fund requirement would reduce giving them an opportunity to use products from the commercial sector that best suits them.
This article has been culled out of my larger treatise on the subject of nearly 8000 words, where an attempt has been made to look at AI in a military context.
BRIEF DESCRIPTION OF ARTIFICIAL INTELLIGENCE
Wikipedia describes artificial intelligence as a machine which mimics human cognition, in doing things like solving problems, though not in the exact same way. Artificial intelligence is the ability of a computer program or machine to think and learn in order to become smart.
Evolution and development of AI systems over time, can help one to broadly understand the different kinds.
Initially, computers handled well defined, simple calculations, that were high volume or high speed tasks, freeing up humans to focus on problems of ever increasing complexity.
In the next level, expert systems were constructed that captured the specialized knowledge of human experts in rules that the system could then apply to situations of interests. This maybe called as the first wave of AI technology, which was quite successful.
This was followed by systems/machines that apply statistical and probabilistic methods to large data sets to create generalized representation that can be applied to future samples for inference.
Then came the deep learning neural networks that can be trained to perform a variety of classification, prediction and generation tasks when adequate historical training data is available. This is the level at which most of the AI is currently placed.
Another interesting way of classifying AI systems, is shown in the table below, adapted from a paper by Andreas Kaplan and Michael Haenlein.
Cognitive Intelligence relates to pattern recognition and systematic thinking; Emotional intelligence relates to adaptability, self-confidence, emotional, self-awareness etc; and Social Intelligence relates to empathy, teamwork, inspirational leadership etc. Artistic intelligence relates to creating something out of imagination, which is currently not present in any artificial intelligence and is unique to human beings.
Analytical AI only has cognitive intelligence, which means that they generate a cognitive representation of the world and use learning based on past experiences to inform their future decisions. Most AI used by firms today falls under this category.
Human Inspired AI has both cognitive and emotional intelligence, which means that apart from all the cognitive features they can also understand human emotions and consider them in decision making.
Humanized AI shows cognitive, emotional and social competencies. Such systems, yet to be built, would be self-conscious and self-aware in their interaction with others so that they actually experience the world in a fundamental way.
DEVELOPMENTS IN ADVANCED MILITARIES
Artificial Intelligence has caught the attention of leaders across the globe and many countries do not want to miss this revolution, unlike the industrial revolution. Countries that had invested in AI long ago have a head start over others but given the political leaders’ emphasis in AI, one never knows who will be the ultimate winner or it may also lead to enhanced cooperation amongst nations, which are aware that given the limited resources, they will be able to achieve their objectives quickly through mutual cooperation.
The position of some of the leading militaries is briefly indicated in the succeeding paragraphs.
United States of America
Though there may have been ups and downs in funding for research in AI over the century, the USA still leads the research in military applications of Artificial Intelligence. Some of their initiatives are:
1. Strategy Document -- The USA Department of Defense has published an AI strategy document, a top level document, which summarizes their intent on harnessing AI to advance their security and prosperity.
2. Joint Artificial Intelligent Centre (JAIC) -- This is the focal point US DOD AI Strategy which is responsible for rapidly delivering AI enabled capabilities to address key missions; establish a common foundation for scaling AI's impact across the DoD; leading all aspects related to data; facilitating AI planning, policy, governance etc; attract and cultivate world class AI team and ensure AI is understood at all levels within the organization.
3. Defense Advanced Research Projects Agency (DARPA)--This agency is responsible for the development of emerging technologies for use by the military and has till now spearheaded large number of programs, especially on robotics, which cannot succeed without AI. There are large number of programs initiated by DARPA covering the entire spectrum of activities for the US Armed Forces, and it is difficult to list them down in this article. A glimpse of some of the latest programs launched by them with strong AI and Robotics inclination, are as listed below:
- AI NEXT. DARPA announced in September 2018 a multi-year investment of more than $2 billion in new and existing programs called the “AI Next” campaign.
- Training AI to Win a Dogfight. DARPA has initiated an Air Combat Evolution (ACE) program that aims to increase warfighter trust in autonomous combat technology by using human-machine collaborative dog fighting as its initial challenge scenario.
- Drone-swarm concept, known as “Gremlins”. The basic concept of operations envisions a C-130-type aircraft as well as combat aircraft, such as fighter jets and bombers, releasing the swarm of Gremlins (Unmanned Autonomous Drones) at a stand-off distance from enemy defenses. 7.
- OFFSET program
The urban warfare presents its own set of challenges to the security agencies who find it difficult to maneuver safely in the congested cities. In order to address these difficulties DARPA has initiated Offensive Swarming Enabled Tactics, (OFFSET).
4. Academia- The US has always promoted a strong relationship between academia and the military and has also benefited considerably from such endeavors to remain at the forefront of technology.
5. Technology companies- It may not written anywhere but the US military always has first access to the latest technologies, which may or may not be released by the technology companies in the USA in the public domain without a nod from US military.
6. Pocket Sized Drones - For the soldiers to detect threats earlier than ever, the US Army has already tested a pocket sized drone, which is small, and has highly maneuverable airborne sensors that are very quiet and difficult to spot.
7. Robots- There are many companies working on robots for defense applications, but one name that stands out is that of Boston Dynamics, spin out of MIT, USA in 1992 (www.bostondynamics.com).
Many developments in China are not immediately or easily visible to the global community and that too in the military and aerospace domains.
However, as quoted by the US defense officials and reported in the open literature, China may leave the US behind in AI arms race. It is understood that China has adopted an approach different from other nations by enabling the technology and the Government sectors work cohesively in "military - civil - fusion".
With a robust emphasis on AI, China is developing next generations autonomous vehicles (drones) in all domains - surface, sub surface and above surface. The famous one being Ziyan's Blowfish - an autonomous above surface drone capable of meeting military missions. Their deep underwater capability was demonstrated when an autonomous vehicle, Qianlong No 2, successfully completed its first exploration in the southwest Indian Ocean for sulfide deposits and precious metals and another one dived and planted China's flag in deep ocean.
Russia is very aggressively pushing AI in all domains, including the military. Their seriousness can be gauged from their President's speech two years ago wherein he had remarked that a country that masters AI will rule the world.
As reported in the open literature, Russia has a large and ambitious autonomous weapons program in place, which it considers is central to its national security interests. Knowing the mathematical prowess of Russians, one would not be surprised, if they come up with innovative AI based war fighting platforms. The cue can be taken from the following events/initiatives:
- The Foundation for Advanced Research Projects (equivalent to USA DARPA) opened the National Centre for the Development of Technology and Basic Elements of Robotics.
- A new city ERA devoted entirely to military innovation is under construction, with the main focus on promoting R & D in military AI and supporting technologies.
- An autonomous drone is being developed which "will be able to take off, accomplish its mission and land without human interference" though weapons use will, for the time being will require human approval.
- Final Experimental Demonstration Object Research, or FEDOR, as it’s popularly referred is another class of robots being developed by Russia. The robot stands six-foot-tall, weighs between 106-160 kg, and can lift up to 20 kgs of cargo. Currently FEDOR is progressing, with its newly acquired ability to hold and fire two pistols at the same time.
The above examples are just tip of the icebergs.
Whilst enhancing the capabilities of weapons with AI on the battle field remains one of the top priorities of any military (some examples have already been mentioned in the previous section, whilst discussing the developments in advanced militaries), there are many background operations and activities that support the man facing the bullets, where AI is being successfully implemented to enhance efficiencies /productivity. The success of AI enabled technologies in the commercial sector are being extrapolated for Defense applications in number of cases, especially for the background activities.
Another aspect that one needs to note is that militaries buy all their equipments and systems from the industry, big or small. For the industry to stay competitive, they are always on the lookout for technologies that will improve the quality and efficacy of their products/systems, thereby providing an edge to the military.
It is through these efforts that one has seen the increase in degree of automation of war fighting platforms, be it tanks, aircrafts, ships, submarines, missiles, smaller and larger weapons, etc, since the days of first world war. The industry is eagerly embracing AI and associated technologies to graduate from automatic systems to autonomous systems.
Some of the applications are listed below:
- Robotic Systems.
- Cyber Security.
- Equipment maintenance.
- Friend or Foe recognition.
- Maps. .
- Predictive policing.
- Simulation and Training.
- Situational Awareness.
- Security Clearance.
The list of AI enabled solutions is endless and is left to one's imagination, provided the militaries overcome some of the following major challenges that they are likely to face in this new emerging technological environment or knowledge revolution:
- Organizational Mindset.
- Organizational Changes.
- Technical Challenges.
- Human Resources.
India was one of the top economies of the world from time immemorial, (one of the reasons of attracting so many invaders), until the industrial revolution, when the British did not allow the intelligent, fertile and innovative mind to come forward and be part of that revolution. The effect of the same could be felt even after independence for many years, until the IT revolution where some pride could be restored in the software sector since the beginning of the 21st century.
Indian Prime Minister's speech at the Defexpo in April 2018 in Chennai sums up the Government's seriousness. He remarked "New and emerging technologies like AI and Robotics will perhaps be the most important determinants of defensive and offensive capabilities for any defence force in the future. India with its leadership in IT domain, will strive to use this technology tilt to its advantage." The Department of Defense Production (DDP) has, in Feb 2018, constituted a task force headed by the Chairman of one of the most respected and reputed company in India, Tata Sons, to study the use of artificial intelligence, which it believes has potential to provide military superiority.
DDP has also directed all the public sector undertakings under their umbrella, to develop and launch AI based products/systems within a year. They are now tying up with academic institutions and others who can offer them solutions in this respect.
DRDO had long ago recognized the potential of AI and accordingly set up a dedicated Centre for Artificial Intelligence and Robotics in Bangalore. They are in turn harnessing the expertise in the academia, industry and startups. It is reported that a ₹73.9 crore project has been sanctioned under CAIR to build AI-based solutions to improve intelligence collation and analysis capabilities of Indian defense. It is understood that CAIR is mulling the development of Multi Agent Robotics Framework (MARF)—to equip the Indian military with a host of robots that could work for surveillance, rescue and mapping operations.
Another project called Energy Harvesting Based Infrared Sensor Network for Automated Human Intrusion Detection was sanctioned for about ₹1.8 crore. Currently, DRDO’s Daksh – electrically powered Remotely Operated Vehicle (ROV) that serve as Bomb Disposal Units for the army, paramilitary forces is already in use.
The Indian Armed Forces are also benefiting from the Niti Aayog's initiatives in AI, under the umbrella of NITI Aayog’s ‘National Strategy for Artificial Intelligence’. Niti Aayog, Intel, and Tata Institute of Fundamental Research (TIFR) joined hands together to set up a Model International Center for Transformative Artificial Intelligence (ICTAI) towards developing and deploying AI-led application-based research projects
The Indian government has also recently announced the National Artificial Intelligence (AI) Portal, developed jointly by the National e-Governance Division of the Ministry of Electronics and Information Technology (MeitY) and the National Association of Software and Service Companies (Nasscom). The portal is meant to work as a “one stop digital platform" for all AI related developments in India.
The AI penetration into the Indian Armed Forces is miniscule, considering the advancements made by other developed nations and our adversaries. The present Government's seriousness can be gauged by the initiatives taken at the topmost level, as mentioned in the earlier section. However, certain recommendations are made in the succeeding paragraphs for the Indian Defense.
National Mission - AI and Robotics
Earmark this as a national mission, wherein all the stakeholders have to pitch in to make this national mission a success, both in AI and Robotics, as they go hand in hand.
The first and the foremost tasks that should be given priority is preparing an AI Strategy document, clearly articulating the vision, missions, objectives and roadmap for the near term and long term goals, by a team drawn from all the stakeholders, especially the end users - Army, Navy and the Air force, as well as the paramilitary forces (since they are the one's who will face both the threats - external or internal, first).
Doctrine for AI warfare
Since there is going to be a paradigm shift in the way future wars are going to be fought once the new technologies are inducted into the military, India needs a new doctrine on the usage of AI/robotics technologies, that will have to flow from the strategy document.
Map the available expertise within the country, who are working in AI and supporting fields, at different levels - the individuals, entrepreneurs, startups, MSMEs, Large enterprises, Academia, R & D organisations etc.
Infrastructure for Data
The backbone and success of any AI venture is data. As AI applications grow in the military, the data storage requirements are going to rise exponentially and hence secure and cyber threat proof systems have to be put in place to store this data at multiple places for redundancy and create robust infrastructure to protect the same. For example, complex neural networks, which are generally data intensive, need to have a vast and flexible storage infrastructure in line with their computing resources. A well thought out and right storage framework is the need of the hour.
Whilst the Indian software expertise has achieved remarkable progress, a very strong emphasis has to be put on the hardware as well. The Government's push in terms of appropriate incentives can accelerate the development of hardware in India.
In this new domain of AI, not very many military personnel have yet been exposed to this new technology or educated or trained. In addition to investing in infrastructure, Indian Armed Forces will also have to focus on getting the right people such as data analysts, data engineers and researchers. They hold the key to their success in AI domain.
Identification of AI applications
At a time when AI and automation cover so many different applications, it can be difficult to know where to focus.
The experts opine to Invest in AI capabilities for tasks with rules or patterns that are predictable and difficult to disrupt. This will give more confidence amongst users. The tasks/processes which are rule based, but performed manually within the defense environment may be taken up first, as the chances of success are large. With success acting as a catalyst, it will encourage users to seek solutions to more complex problems as well.
The tasks with indescribable or unpredictable rules could be taken up later.
Human in the loop
Whilst there are still debates ongoing about whether to make weapons systems fully autonomous or keep humans in the loop, in a military environment where there are no runners up, the author is of the view that initially, it is essential to have human in the loop where routine number crunching is preformed with AI that supports quicker decision making, especially in an unpredictable world.
As India is on a mission to embrace AI as quickly as possible and with anyone and everyone talking about AI, it is important for the Armed Forces personnel to have an understanding of AI, ML, Robotics etc. The depth and breadth of such an understanding will vary, depending on the person's hierarchy, and his involvement with AI enabled products/systems - at design, operational or maintenance level. Appropriate training modules could be embedded in different courses all over the training institute.
AI is the most essential element in autonomous robots which are vehicles for AI to function/perform efficiently. This field also requires equal, if not more attention, so as to draw the full benefits of AI. It is finally the robotic soldier that is going to face the bullet in the battlefield and if a nation has built this capability, it will save precious human lives.
R & D Funding in AI
The Armed Forces should earmark a separate fund, and operate differently from the routine acquisition process using DPP or DPM, for AI based research at all levels - Basic and Fundamental as well as Applied. The author has been advocating a different set of guidelines if we want to leverage on the innovative capabilities of an Indian, that is by having a "Defense Indigenous Systems Development Guidelines" that gives more autonomy to the head of organization to fund projects. The research areas may include data management; maintenance; logistics; cyber resiliency; platform safety; etc.
Large investments are being made in the field of AI and Robotics by different nations to gain an upper hand in the war. AI technologies are at different levels of maturity in different countries as well as in different domains and applications within the Defense. The fact that there exists a huge potential for these technologies to play a major role has already been demonstrated by some of the private companies in a very short time, points to a time not too long in the future when the nature of warfare itself will undergo a transformational change. It may never be the same as experienced in the written human history.
Data is everything in machine learning, which means whoever gets access to data can have a big advantage. The Indian Government has therefore justifiably commenced enforcing data localization for all the companies, be it foreign or Indian, despite their opposition.
India has to take a leap forward in developing autonomous systems, before Lethal Autonomous Weapons Systems (LAWS) discussions at the UN concludes. It is similar to nuclear weapons development, when few nations, after having acquired this capability, prevented others from developing such a capability.
It is encouraging to note that the Indian Government is rightfully very serious about Artificial Intelligence and has fully recognized that AI capabilities need to be leveraged for the larger benefits of all the stake holders - from delivering process efficiencies through back office process automation, to supporting improved decision-making to enhancing military capabilities.
The author wishes to acknowledge the support given by Mr Rahul Rana in preparing and reading through the complete paper.
Commodore (Dr) R K Rana, an Indian Navy veteran with 33 years of an illustrious career on board ships, naval dockyard, training, research, staff duties, design of warships and indigenous products development organisation. He is presently, a Honorary Senior Advisor at the Foundation for Innovation and Technology Transfer of IIT Delhi.