(This article was published in SYNERGY : Journal of the Centre for Joint Warfare Studies (CENJOWS) Special issue on Impact of Future Technologies on Warfare, August 2017)

The epic image of the Terminator standing on a mound of human skulls in James Cameron’s sci-fi film Terminator-2 conjures up a future in which robots may fight with and/or against their creators – humans. Mind-boggling advances in robotics, machine learning, and data analytics are transforming modern and future warfare. Future wars will be fought with military robots, or Autonomous Weapons Systems (AWS), defined by the U.S. Department of Defence¹ as “systems that function with minimal to no human intervention, selecting and engaging targets by themselves through pattern-recognition algorithms.”

Modern warfare may see the use of covert swarms of miniature spy drones, unmanned ground vehicles, missiles with decision-making powers, sentry robots, and offensive robot strikers. At the centre of it all is Artificial Intelligence (AI), or computer algorithms that can perform many functions, such as vision, decision-making, and the ability to process vast quantities of information, capabilities normally associated with humans. Effective use of AI leads vehicles and weapons to identify and attack targets automatically through pattern recognition and face recognition algorithms. The U.S., China, Russia, Britain, and other developed countries are involved in harnessing AI and weaponizing it. The US has put AI at the centre of its quest to maintain its military dominance and has recently appointed Google’s Eric Schmidt as the chairman of a new Defence Innovation Advisory Board².

State of Technology and Current Weapon Systems

Military robots are autonomous or remote-controlled robots designed for military applications, from transport to search and rescue to attack. Most countries have adopted near-autonomous systems in protective mode as defensive methods against incoming threats. Defence systems like the Iron Dome of Israel³ and the Phalanx Close In Weapon System⁴ of the U.S. Navy are examples of near-autonomous defence systems. South Korea uses an automatic sentry robot called SGR-1⁵ against inbound threats in its Demilitarised Zone with North Korea.

Drones heralded the era of machines that could be operated remotely, thousands of miles away from the safe confines of war rooms. The drones can perform reconnaissance without endangering a pilot’s life and cost a fraction of the cost of manned aircraft. The U.S. has, in fact, used them for targeted killings in Iraq, Yemen, Afghanistan, and most recently Pakistan. Israel has also successfully employed them against its enemies. Modern military drones, like the MQ-1B Predator⁶, can loiter over enemy territory for more than 14 hours and perform precise surveillance, including HD video, still photography, and thermal imaging. The infusion of AI into weapon-carrying drones can make them lethal flying autonomous systems that decide their targets themselves. Taranis⁷, the joint Anglo-French drone programme, has capabilities for autonomous strikes and will be ready as early as 2030.

In addition to drones, the US armed forces already employ a host of robotic systems, from bomb disposal robots to reconnaissance and attack drones. The US is reported to have tested missiles that can dynamically decide their targets and built ships that can hunt for enemy submarines without human intervention. Not to be left behind, China is incorporating AI technologies into a new generation of anti-ship missiles⁸ with autonomous targeting capabilities in response to the “semi-autonomous” Long Range Anti-Ship Missile expected to be deployed by the US in 2018.

Towards the end of the last year, Strategic Capabilities Office⁹, U.S. Dept. of Defence, demonstrated the collective decision-making and adaptive formation flight capabilities of a swarm consisting of 103 Perdix drones launched from three F/A-18s. It may be considered a major milestone in the development of small, inexpensive, and autonomous weapons systems. Such swarms don’t consist of pre-programmed individuals; instead, they participate in collectively analysing situations and adapting to them, much like swarms of bees and schools of fish. Present-day anti-air defences and radar systems are simply incapable of detecting things this small. A swarm of autonomous drones would fly into a combat zone to jam enemy communications, provide real-time surveillance capabilities, and autonomously fire against the enemy when threatened.

Terminator Conundrum

The debate within the military community is no longer about whether to build autonomous weapons but how much autonomy to give them so that they could decide whom and when to kill. It’s something they call the “Terminator Conundrum¹⁰”. Autonomous weapons that target objects like missiles, tanks, and ships are less contentious. However, autonomous offensive weapons that can kill humans raise serious concerns about their legality and adherence to laws of modern conflict.

AI robots may be free of human errors, but capable of catastrophic mistakes of their own. Autonomous robots with the ability to make life-or-death decisions and out-kill the enemy pose some questions before all of us. Whose fault it will be if a robot causes the wrongful death of a civilian – commander, programmer, manufacturer, or the robot himself? Is it ethical to let machines handle our killing for us? Can the robots be tried for war crimes? Is getting killed by a robot more humiliating than by a fellow human being?

These questions don’t have easy answers.

The Automatons

‘Autonomous’ systems are way more advanced than the automatic’ or ‘automated’ systems. Report¹¹ of the UN Special Rapporteur on the extrajudicial, summary, or arbitrary executions states that “Automatic systems, such as household appliances, operate within a structured and predictable environment,” and “autonomous systems can function in an open environment, under unstructured and dynamic circumstances.” Robotic weapons, which are autonomous, are divided into three categories based on the amount of human involvement:

Human-in-the-Loop Weapons: Robots that can identify targets and attack them only with a human command;
Human-on-the-Loop Weapons: Robots that can identify targets and launch attacks under the supervision of a human operator who can override the robots’ actions; and
Human-out–of-the-Loop Weapons: Robots that are capable of identifying targets and launching attacks without any human input or interaction.

The majority of current military robots are tele-operated and not weapon-equipped; they are employed primarily for reconnaissance, surveillance, disposing of explosive devices, etc. Most military UAVs are only slightly more advanced: they have some low-level autonomy that allows them to navigate, without human intervention, but almost all require significant human involvement to complete their missions.

The ever-changing battle conditions call for “human judgement, understanding why people do what they do, anticipation of the results of actions, and appreciation of the larger picture.” While AI-powered robots can be superior to humans in some rule-based tasks, in uncertain battle situations that require judgement and understanding, humans are way superior to robots. The holy grail of trustworthy autonomous weapons is developing systems whose behaviour is predictable enough so that they can be safely deployed, yet flexible enough that they can handle fluid war situations.

What we stand to gain

Autonomous robots can save soldiers’ lives by removing soldiers, who might otherwise get killed, from the battlefield. Their speed and precision will be unmatched by any human being. The intent is clear: to plug the gaps between physical endurance and human thinking using machines. Major Kenneth Rose of the US Army’s Training and Doctrine Command¹² outlined some of the advantages of robotic technology in warfare:

“Machines don’t get tired. They don’t close their eyes. They don’t hide under trees when it rains, and they don’t talk to their friends … A human’s attention to detail on guard duty drops dramatically in the first 30 minutes … Machines know no fear.”

Autonomous machines, devoid of human emotions, will comply with the laws of war (LOW) and rules of engagement (ROE). Since AI-powered machines are unaffected by fatigue, mental stress, and emotions, they are unlikely to make the rash decisions common on emotionally charged battlefields. Manpower-intensive tasks like demining, patrolling, and providing logistics are mundane and sometimes dangerous too. Autonomous weapon systems that are unaffected by mental stress, fatigue, and boredom can perform such routine tasks for longer durations. In combat situations, such a system can rapidly analyse data inputs in an objective manner and take decisions based on analysis and not out of panic.

What we risk to lose

Extensive use of autonomous systems will significantly reduce the number of human casualties on the battlefield, thus reducing the human cost of war. This may prompt countries to go to war more readily rather than explore other avenues of conflict resolution.

NGOs and think tanks such as Human Rights Watch¹³ and the Campaign to Stop Killer Robots¹⁴ have been urging world powers and the United Nations to outlaw the development of “Lethal Autonomous Weapons Systems (LAWS)”. The Future of Life Institute¹⁵ published an open letter in July 2015 signed by thousands of AI researchers and famous tech personalities like Stephen Hawking and Elon Musk warning, “If any major military power pushes ahead with AI weapon development, a global arms race is virtually inevitable, and the endpoint of this technological trajectory is obvious: autonomous weapons will become the Kalashnikovs of tomorrow.” The open letter calls for an immediate moratorium on the development of such systems because their easy availability and ability to be used thousands of miles away will increase the likelihood of armed conflicts. Further, signal jamming and computer hacking could disrupt transmission, hijack autonomous weapons, and turn them against their operators.

The United Kingdom and the United States do not support an outright ban, given the benefits of efficiency and fewer casualties; however, they have agreed that regulation may be needed. A complete ban will halt all research activities related to any kind of defence automation. The Ottawa Treaty is often mentioned as the precedent for a pre-emptive ban, but major powers like India, Russia, China, and the United States are not signatories to the treaty, making it almost ineffective. Any ban will be effective only if supported and upheld by all major powers, and it seems highly unlikely in the near future.

Why India needs them

The issue of autonomous weapons in India should be seen in the light of its precarious security situation: two hostile neighbours, the ever-present threat of insurgency, and naxalism. The use of autonomous weapons will have multiple benefits, including reducing casualties, improving surveillance, using a smaller but lethal force, and faster deployment. Their greatest benefit is that they are not unaffected by fatigue, stress, or emotions; hence, they can be deployed 24×7, 365 days a year, in harsh weather conditions. Manned patrols are very tedious and risky along the LoC and in the naxal-infected tribal areas. Indian soldiers are regularly exposed to the vagaries of nature in Kashmir Valley and Siachen Glacier, the highest battlefield in the world, resulting in heavy casualties due to avalanches, frost bites, and high altitude sickness. Autonomous robots can do a much better job of carrying out surveillance and patrolling in such areas.

India has one of the biggest constellations of satellites placed in space, and it has come under great threat after a successful Chinese anti-sat test in 2007¹⁶. An autonomous defence system that detects incoming threats to satellites and launches missiles to intercept them would deter the plans of hostile neighbours. Also, AI might bridge the ever-increasing gap between the conventional military capabilities of India and China.

Autonomous Weapons in Indian Arsenal

The Centre for Artificial Intelligence (CAIR), a DRDO laboratory, is developing a wide array of robotic systems for reconnaissance and surveillance¹⁷. These include RoboSen, a mobile robot system targeted at patrolling, reconnaissance, and surveillance applications for the armed forces, paramilitary agencies, and police. It is capable of autonomous navigation in semi-structured environments with obstacle avoidance capabilities and continuous video feedback. It has also developed a ruggedized man-portable Unmanned Ground Vehicle (UGV) for low-intensity conflicts and surveillance in urban scenarios and a ‘Snake Robot’, fitted with a high-definition camera and ultrasonic sensors, flexible enough to navigate difficult passages and look for survivors. CAIR has also indigenously developed walking robots with six and four legs for logistics support, as well as wall climbing and flapping wings for potential usage in Low-Intensity Combat (LIC) operations.

Having developed these advanced robots, CAIR is now in the process of developing the Multi Agent Robotics Framework (MARF)¹⁸ – with a host of robots collaborating as a team for carrying out surveillance, area exploration, search, rescue, etc. Ensuring the mobility of such systems in a variety of terrains and environments—mountainous, desert, rural, urban, outdoor, and indoor—is still a locomotive challenge. Another DRDO laboratory, the Research and Development Establishment (Engineers) has developed Daksh. The robot is primarily designed to detect and recover IEDs and now has a CBRN detection mechanism¹⁹.

India’s Position on Autonomous Weapons

At the United Nations Convention on Conventional Weapons (CCW) Meeting of Experts on LAWS, India stated²⁰ that “there is a need for increased systemic controls on international armed conflict in a manner that does not widen the technology gap amongst states or encourage the increased resort to military force in the expectation of lesser casualties or that use of force can be shielded from the dictates of public conscience.” India has also expressed its concerns about the proliferation of such weapon systems, particularly by non-state actors.

The way ahead

Though India has made a few strides in the area of autonomous weapons, it is still a long journey. The following steps can be taken to speed up the process:

Advances in AI inside military systems rely largely on the capabilities of engineers and programmers to design intelligent systems that demonstrate human-like reasoning and pragmatic decision-making. Such developers should have expertise in software, hardware, robotics, etc. Currently, there are not many institutes (barring a few IITs) that have undergraduate and postgraduate courses in the fields of mechatronics, controls, robotics, and machine learning. More institutes should be founded and existing ones upgraded to cater to these engineering disciplines.
Long-term investments should be made in AI research, the present budget allocations are simply not enough.
India has a thriving software industry, and greater collaboration between military establishments and I.T. companies can immensely help.
AI should be made a critical component of the Make in India, Skill India, and Digital India programmes by offering incentives for manufacturers and creating clusters between industries, start-ups, and educational institutes.

Conclusion

Despite the international outcry against the use of robots in modern warfare, several technologically advanced countries have already undertaken the development of AI-based killer machines. India, too, must recognise the benefits of such systems in terms of lives saved given the country’s naxal problem and Pakistan-sponsored proxy war. Autonomous weapons will augment India’s defences and may, in fact, be better able than human soldiers for tasks like checking cross-border infiltration, patrolling, and reconnaissance. Internationally, India must voice its concern about the regulation of the design, development, and use of autonomous systems due to their potential misuse. Since a complete ban is an unachievable utopia, pragmatic and non-lethal deployment of such systems should be India’s stand. Indigenous development of autonomous weapons should be an integral part of the Indian Armed Forces’ modernization plans, and it is high time that the country puts in efforts, resources, and skilled manpower to pursue them. In the end, “The more you sweat in research, The less you will bleed in war.”

REFERENCES

Department of Defense, DIRECTIVE, NUMBER 3000.09, 21 November 2012
“Former Google CEO Schmidt to head new Pentagon innovation board”, Reuters, 2 Mar 2016.
“Israel successfully tests improved version of Iron Dome”, The Times of Israel, 23 February 2017.
Phalanx Close-In Weapon System; available at http://www.raytheon.com/capabilities/products/phalanx
“Who goes there? Samsung unveils robot sentry that can kill from two miles away”, Daily Mail UK, 15 September 2014.
MQ-1B Predator available at http://www.military.com/equipment/mq-1b-predator
“This drone is one of the most secretive weapons in the world”, Business Insider India, 30 Sep 2015.
“Terrifying Future of war: China’s next cruise missiles will use ARTIFICIAL INTELLIGENCE”, Express UK, 19 Aug 2016.
“The Pentagon Launches 103 Micro-Drones In World’s Largest Test Of ‘Swarm’ Technology Yet”, indiatimes.com, 10 January 2017.
“The Pentagon’s ‘Terminator Conundrum’: Robots That Could Kill on Their Own”, The New York Times, 25 Oct 2016.
Report of the Special Rapporteur on extrajudicial, summary or arbitrary executions, Christof Heyns, 9 April 2013.
“Robot soldiers”, BBC News, 12 April 2002.
“Killer Robots”; available at http://www.hrw.org/topic/arms/killer-robots.
Campaign to Stop Killer Robots; available at http://www.stopkillerrobots.org.
“An Open Letter : Research Priorities for Robust and Beneficial Artificial Intelligence”; available at https://futureoflife.org/ai-open-letter
“Flexing Muscle, China Destroys Satellite in Test”, The New York Times, 19 Jan 2007.
CAIR website available at https://www.drdo.gov.in/drdo/labs1/CAIR/English/indexnew.jsp?pg=products.jsp
“Army to get self-reliant, autonomous robots soon”, Times of India 4 Mar 2017.
“Faster, lighter, DRDO’s Daksh now has CBRN detection mechanism”, The Indian Express, 26 May 2015.
Statement delivered by PR to CD at Plenary of the fifth review conference of the CCW on 1December 12, 2016; available at http://meaindia.nic.in/cdgeneva/?5677?000

Short Stories

The Hold

Short Stories

The Hold