The Wavell Room
Image default
Automation and Technology Concepts and Doctrine Long Read

What implications will the ongoing developments in artificial intelligence and robotics have for remote warfare in the future?

Forward – Vernon Coaker MP former Shadow Secretary of State for Defence

The increasing use of robotics by our armed forces is inevitable in all areas. The key question for the military but also the public is how do we ensure that the artificial intelligence that goes with it will act according to an acceptable moral compass. This will also be a challenge for the military operators of such technology and for political oversight. In particular how do the existing rules of war with the various conventions apply to cyber attack. For example, does Article 5 of NATO apply to an attack on the computers which control water or electricity supply. These are just some of the questions which arise as well as many more, but it is ever more important that this discussion takes place with the wider public as well as the senior military and politicians. A failure to do so will undermine the technological advances in warfare which are both inevitable and necessary.

Executive Summary

The character of warfare changed with the onset of the 21st century. The decade prior to this saw warfare predominantly characterised by intra-state conflict drawn across ethnic lines, often resulting in humanitarian crises. This led to the development of western military forces being utilised for stability operations (Aoi 2011); ranging in nature from peace enforcement in Somalia in 1993, peacekeeping in Bosnia in 1995, to humanitarian assistance in Kosovo in 1999. The events of 9/11 changed this, ushering in the United States (US)-led Global War on Terror (GWoT). What started as an initially limited intervention in Afghanistan to dismantle the Al-Qaeda training camps and topple the rogue Taliban regime in Kabul, soon developed into the 2003 invasion of Iraq, with subsequent interventions across the Middle East and Africa (Libya 2011 (1), Mali 2013 {2), Nigeria 2015 (3), and Cameroon 2015 (4) amongst others). Often framed under the narrative of countering international Islamic terrorism and the GWoT, these long and costly counterinsurgencies across the wider Middle East and Sub-Saharan Africa, and in particular the wars in Afghanistan and Iraq (Bankoff 2003), fundamentally altered the character of war, heralding an era of remote warfare (RW), defined by western governments which are politically averse to military risk when pursuing strategic objectives.

his paper will explore how the nature of RW has evolved over the last twenty years since 9/11, moderately steady in some areas yet alarmingly high in others. By deconstructing precisely what is meant by RW its respective component parts can be subsequently scrutinised individually as part of RW’s evolutionary nature. Once these component parts of RW have been established, it will be necessary to place this within current operational theatres to give a wider understanding of how RW is changing the nature of war. In addition, the United Kingdom’s (UK) approach to RW will need to be explored, and in that context to investigate the UK military’s current policies towards artificial intelligence (AI) and robotics; arguably the area of RW which is currently seeing the most rapid developments with the emergence of robotics technology and more advanced AI platforms being pressed into service. Once this approach has been established within the current international environment, it will be possible to elucidate further the implications for the future of RW when considering the ongoing developments in both AI and robotics.

The evolving characteristics of Remote Warfare

Remote warfare, as both a concept and as a continuously evolving character of (predominantly) westernised warfare, has seen relatively little debate, both in terms of policy implementation and subsequent delivery5, and also little within the wider academic discourse. Rather, the individual component parts which constitute RW as a means by which a state can achieve military success have received much wider scrutiny, both contextually (Smith 2015) and operationally (Gross 2015). RW constitutes a strategy of countering military threats at a geographical distance, negating the requirement for a traditional deployment of a potentially large military force. In a political climate whereby both cost-effective policies combined with reduced military risk often equates to the preferred policy options, there is a case to be made that a political precedent was set during the NATO campaign in Kosovo regarding the utility of RW. Much of the decision to use air power during Kosovo was due to risk aversion across NATO and specifically from then-US president Clinton when Tony Blair and others believed that a ground force was a good option. Constraints placed on air operations, specifically missions conducted at high altitude to avoid Surface to Air missile threats, reflected these concerns and a culture of risk aversion. Arguably air power represents lower risk and political commitment than ground troops; RW is a further evolution of this. Specific RW policies have already altered, remarkably so, in the time period between the emergence of the GWoT and the contemporary international environment. In order for a contextual appreciation of where AI and robotics lie within the conceptional space of modern warfare, the various segments which make up RW must first be established and analysed. These methods, discussed below briefly, include the use of unmanned ariel vehicles (UAVs), train and equip programmes of foreign militaries, the increased propensity to rely on Special Operations Forces (SOF), and the amplified expectation of host nations in a conflict to undertake the majority of the ground close combat.

The use of UAVs in the conflicts since the invasions of Afghanistan and Iraq serves a direct requirement for the ability to project force remotely and reduce the risk to both pilots and human controllers, who are often not in the conflict zone itself, and their ground crew and

support, who are often in another country entirely, far removed from any physical threat. The development of UAV technology and subsequent operational use by the UK and in particular the US has grown exponentially over the last decade. While there were a total of nine strikes from UAVs between 2004 and 2007, figures released by the USAF and the RAF showed that there had been over 1,400 UAV airstrikes in Afghanistan between 2008-13, with armed UAVs carrying out a quarter of all air combat air sorties within Afghanistan towards the end of the combat phase of the conflict6. By 2013 the US had over 240 armed UAVs in operation7, and was known to have operated in at least six countries; Iraq, Afghanistan, Pakistan, Yemen, Somalia and Libya. More recently, in current operations in Iraq and Syria against Islamic State (IS) the UK has conducted more than 1,925 UAV sorties resulting in 4,100 strikes8. This represents an increased utility over a similar period during the height of the conflict in Afghanistan. The evolving character of remote warfare in this instance is highlighted by the recent announcement that UK military personnel involved in the fight against IS in Iraq and Syria who were not physically in the countries will for the first time be eligible for the campaign medal; specifically, UAV pilots and support personnel who are often based either at RAF Akrotiri in Cyprus or RAF Waddington, Lincolnshire. This highlights the increased capability of UAV technology in finding, identifying, fixing, and striking enemy combatants remotely, using AI. UK Defence Secretary Gavin Williamson stated that; The expanded medal criteria means that those personnel who have played a vital role in defeating Daesh but have been based outside the conventional area of operations will receive the recognition they deserve”, and that this; “reflects the changing character of warfare”9.

The UK witnessed a shift in strategy from the post-Cold War period to the emergence of the GWoT with regards to train and equip programmes. Whilst committed in overseas interventions between 1991 – 2001, notably the first Gulf War, Bosnia, Kosovo, Macedonia and East Timor, the military never really developed train and equip programmes during these conflicts for their respective militaries. However, this changed significantly with both Afghanistan and Iraq. The Iraqi Army (IA), once rebuilt in the wake of the Coalition Provincial Authority’s disarming, was trained and equipped by both the US and the UK militaries10. Indeed, the priority for the British Army towards the end of the campaign, and a central pillar of its strategy throughout, was to train the IA to an operationally sufficient standard so as to take the lead in security operations across southern Iraq. This was achieved by 2008 with the

IA-led operation Charge of the Knights11 in Basra. Small pockets of platoon-sized units, known as Military Transition Training teams (MiTTs) remained embedded within IA battalion-sized locations throughout Basra city until the end of the British campaign in 2009, in addition to the Naval Training Teams (NaTTs) based out of Um Qasr, training and equipping the Iraqi naval forces12. The same is true of the Afghan National Army (ANA) in the immediate aftermath of the Taliban’s fall from power in 200113. Indeed, scaling up and developing operationally a host nation’s own military has been a fundamental objective in western campaigns since 9/11. A metric by which this is deemed either successful or not is both the quantitative increase in the troops trained, in addition to their relative qualitative operational output. Once the UK had completed its combat operations in Afghanistan by 2014, the emphasis for the remaining military commitment returned to training the ANA; this time around its officers and leadership class in Kabul. By 2017, UK military trainers had trained and equipped 3,000 ANA officers14. The train and equip programmes run by the UK have to a large extent been successful, producing well trained professional soldiers and officers in both Afghanistan and Iraq.

Whilst just as central to these campaigns, state specific train and equip programmes have been utilised extensively by the US since first authorised under the 2006 Defense Bill to conduct limited military action overseas in the name of fighting international terrorism15. Whilst the initial figure for these programmes was US$200 million in 200616, this rose significantly to US$2.5 billion for the fiscal years 2016/1717. Despite such a steep increase in available resources, the train and equip programmes have recently received internal scrutiny highlighting concern over structural issues. In 2017, the US Department of Defense Inspector General conducted a thorough appraisal of the US train and equip programmes, finding a lack of overall strategy in implementation of the policy, in addition to failing to adequately staff these programmes to match the generous budgetary increases18, leading to operational shortcomings and waste. Certainly, whilst the doctrine of train and equip makes robust policy sense from a RW approach, a higher degree of external validity and oversight would ensure for greater success in the future of RW. This becomes especially evident when combining the train and equip programmes with enabling host nations to conduct the majority of the ground close combat, discussed further below.

A further characteristic of RW is the often-heavy reliance on SOF personnel. The ability for a military to dispatch highly specialised SOF units into a low or medium level intensity conflict area often negates the requirement for sending in much higher numbers of conventional armed forces. This policy option has proved highly popular in the west, particularly the US. Whilst the battles and combat in Iraq and Afghanistan were arguably won by the infantry, the period of the GWoT was again the defining moment in modern history whereby the capabilities of the SOF were dramatically increased, both in manpower and resources. The US and the UK both have long histories of utilising SOF units in their numerous military campaigns throughout the twentieth century. However, this was primarily in a counter-insurgency role (Malaya, Oman, Vietnam, Central America). The first shift in SOF utility occurred towards the end of the twentieth century, with two major international state-based conflicts; the British Falklands campaign in 1982, and the US-led Gulf War in 1990/91. During these campaigns, small units of men, often deployed far behind enemy lines, were used to augment the infantry on the frontline against a relatively professional army (defined by paid volunteers / conscripts; the wearing of a recognisable military uniform to differentiate from civilians; and bound by some extent to the international laws of armed conflict). These were the last great state-based conflicts in modern times. The warfare of the Global War on Terror a decade on from the Gulf War was characterised by a return to counter-insurgency; in the wake of the collapse of both the Afghan and Iraqi military and police establishments, this necessitated the defence of the population against insurgent groups determined to bring about a change in government not by the ballot but by violence. With such a change in warfare and the return to counter- insurgency, another shift in the utility of SOF capabilities and deployments was witnessed.

Understandably the nature of SOF operations warrants a certain degree of discretion in order to protect operational security. Therefore, the availability of data for the purpose of analysis is often delayed. In 2016, US SOF units had deployed to 138 states, roughly three quarters of the globe19, with approximately 8,000 SOF personnel deployed. This figure has more than doubled since 2001, when the number of deployed personnel was 2,90020. The presidency of Barack Obama oversaw this operational shift; reducing conventional forces in conflict zones from 150,000 in 2008 to 14,000 in 201621, yet SOF forces remained consistent. Their budget increased also, from US$9.3 billon to US$10.4 billion, a significant 10% increase, in addition to 15,000 further personnel. This development of SOF capability increased further still under President Trump; soon after taking office he authorised an increased remit for personnel

operating in Yemen, with at least 92 strikes conducted in the first half of 201722. Data from casualty figures can illustrate further the expanding roles of SOF personnel. Despite accounting for 5% of US forces, since 2016 SOF personnel have accounted for more casualties than have conventional personnel. This change was reinforced in 2017 by the deaths of SOF personnel in Yemen, Somalia and Niger. The deaths of four personnel in a single incident in Niger highlighted the strategic significance that the US (and, to an equal extent, both the UK and France) place upon Africa. In 2006, only 1% of SOF personnel were operating in Africa. By 2017, that figure had climbed to 17%23, representing the policy challenges for politicians eager to engage in intervention yet all too aware of the political risks that conventional forces pose. The degree to which SOF operations have increased over the last decade in particular has been highly significant as an example of how not only RW has evolved throughout the period, but also of the extent to which it has become the preferred policy for politicians wishing to engage in small to medium level intensity conflicts, as opposed to the deployment of conventional military forces.

The increased reliance on SOF personnel to undertake operations, in addition to their role in train and equip programmes, is key to RW as a concept, enabling intervening nations to limit their direct involvement whilst still achieving a strategic effect. Central to this approach is the operational shift from intervening nations’ military forces to host nations’ forces involved in most of the combat, incorporating both the increased utility of SOF personnel with the train and equip programmes discussed above. In contemporary conflicts this has been the case across the Middle East in the fight against IS; US-led forces have trained both the Iraqi Army and the Iraqi Peshmerga, and the Syrian Democratic Forces in Syria, to identify the two most significant training and advisory missions in recent times, whereby the host nations have conducted the majority of the fighting. To indicate the scale of these missions, in Iraq, as of December 2016, there were approximately 15,000 US forces alone, two thirds of which were directly involved in training Iraqi security forces. 65,000 of the latter were calculated to have benefited from this program24. In Syria, US equipped and trained allies the Syrian Democratic Forces (SDF), and their majority component the People’s Protection Units (YPG), have been involved in the majority of offensive military operations to retake territory claimed by IS across northern and eastern Syria. Despite maintaining no fewer than 2,000 military personnel in Syria involved in training Syrian opposition groups, and potentially up to at least double that25,

the US has left the majority of the fighting to the SDF and others. This is highlighted by the respective casualty figures for ground close combat against IS in Syria in 2017; the SDF suffered 955 fatalities26, whilst the US suffered zero casualties that year. However, weeks after President Trump’s announcement of a US withdrawal from Syria in December 2018, four US personnel were killed in action in a single incident in Manbij, northern Syria27; previously there had only been two combat-related fatalities for US forces in Syria since 2015.

Despite the clear advantages to policy makers in having a host nation conduct the majority of the combat, in addition to employing SOF personnel to conduct the train and equip programs, serious questions remain as to the feasibility of not only maintaining this current trend in future conflicts, but also the significant upward trajectory of its utility as policy. The issues arise largely as a result of choosing who the local actor will be, along with any issues or discrepancies with their operational integrity. Problems which come to develop here, for instance, will be seized upon as evidence of poor prior strategic insight. In Syria, though the YPG have performed well on the battlefield against IS and other Islamist forces, there have been recurring issues relating to war crimes and human rights abuses suspected to have been committed throughout their actions in the conflict (Federici 2015). Often local actors, even host nation militaries operating in developing regions, are either unwilling or unbound by the same international laws of armed conflict which western professional militaries are. In addition, local actors may become either too powerful, or have questionable allies or motives themselves. For instance in Syria, a proportion of US-trained rebels in the early stages of the conflict handed over their weapons to Jabhat al-Nusra, an Islamist terrorist organisation linked to al- Qaeda (AQ)28. Further still, the YPG are viewed by Turkey as terrorists29, and indeed share the same ideology, command structure and financing as the PKK, a proscribed terrorist organisation by the US, UK and EU. This all leaves the question of sustainability once the mission is completed. In the aftermath of the previously announced US withdrawal from Syria, likely to be completed largely by the end of 2019, what fate will befall their Kurdish allies, and how will they react with other actors to the ensuing power vacuum? In the wake of the US forces’ withdrawal, Turkey has already made clear its intent to take the northern city of Manbij, seen as a trouble-spot for Kurdish separatist and terrorist movements inside Turkey. This could likely result in NATO’s second largest military being occupied with fighting a counter- insurgency directly on Europe’s and NATO’s southern boundary, tying up both Turkish

resources and manpower away from other NATO tasks. Highly significant long-term impacts can have a highly destabilising local and regional effect if the utility of local actors is not managed to a high degree, not just during the campaign, but crucially also afterwards, resources and practicalities permitting.

The rise of AI & Robotics within Remote Warfare

In addition to the increased use of train and equip programs, of SOF personnel, and in host nations taking the dominant role in ground close combat, RW as a concept also covers the very broad and often multi-dimensional use of AI software and robotics technology in military equipment. This is a further progression by policy makers seeking to limit both loss of human life whilst conducting conventional warfare in addition to, it is hoped, long-term cost saving by increasingly incorporating technology to assist in and, in certain circumstances, relieve human action. Whilst the incorporation of AI and robotics into militarised technology is not a recent phenomenon, the pace of recent development combined with the speed to market of these increasingly emerging technologies does certainly require pause for consideration. The militaries of the US, China, Russia, Iran and increasingly the UK are seeking to devote considerable sums to upgrade their current AI and robotics platforms whilst seeking additional large investments to develop the next generation of AI capability. As part of the 2018 Defence Modernisation Programme, the UK Ministry of Defence announced £160 million for the Defence Transformation Fund, investing in new concepts and technologies aimed at improving AI technologies30, which will potentially rise to £500 million later this year. Also in 2018 the Pentagon established the Joint Artificial Intelligence Centre (JAIC), set to coordinate all 600 AI projects within the US Department of Defense, with a budget of $US1.7 billion31. Then-US Deputy Secretary of Defence Shanahan stated that this restructuring was a departmental priority, emphasising the requirement for a swift conclusion32. This investment in new technology is replicated globally; total global spending on robotics (including unmanned aerial vehicles) will likely total US$115.7 billion in 2019, a 17.6% increase over 2018 spending. This figure is expected to reach US$210.3 billion by 202233, representing a substantial increase in both AI and robotics which will impact significantly the short-term future and evolving nature of RW as it becomes further integrated.

At the base scientific and technology level, advances in machine autonomy derive primarily from research efforts in three disciplines: artificial intelligence (AI), robotics, and control theory. The US is the country that has demonstrated the most visible, comprehensive and perhaps successful military research and development (R&D) efforts on autonomy. China and the majority of the nine other largest arms-producing countries have identified AI and robotics as important R&D areas. Several of these countries are tentatively following in the US’ footsteps and looking to conduct R&D projects focused on autonomy, in particular China, Russia, Iran and the UK. Civilian industry leads innovation in autonomous technologies; as expected, Silicone Valley as the tech capital of world pioneers some of the most advanced R&D, particularly Google. Other influential players are major information technology companies such as Alphabet (Google), Amazon and Baidu, and large automotive manufacturers including Toyota that have moved into the self-driving car business. This civilian-developed technology is being utilised increasingly for military purposes, the Pentagon stating in 2018 that the military will acquire driverless vehicles before the public34. In addition, traditional arms producers are certainly involved in the development of autonomous technologies, but the amount of resources that these companies can allocate to R&D is far less than that mobilised by large commercial entities in the civilian sector. However, the role of defence companies remains crucial, because commercial autonomous technologies can rarely be adopted by the military without modifications and companies in the civilian sector often have little interest in pursuing military contracts.

In order to analyse the future implications for increased and sustained use of AI and robotics, it will be necessary to determine the current roles AI is utilised in and its recent developments. Due to US AI programmes often having far larger resources made available to them, this will be from a US perspective in addition to a UK approach, in order to assist with determining where AI is heading within the military. There are three classifications of AI technology; narrow, general and super. Artificial Narrow Intelligence (ANI) is technology which has tightly controlled parameters for operating and is within a specific context. This technology is already in use for example Siri by Apple and RankBrain by Google. Artificial General Intelligence (AGI) is where a system can operate at the same level as that of a human. Most experts believe that this may be possible, but that it is still a considerable way off, potentially as far as at least204035. Artificial Super Intelligence (ASI) is whereby platforms can supersede both human intelligence and control, believed to be at least a century away from achievement36.

Currently, no military in the world is capable of developing existing platforms beyond ANI, and those already in use are set restrictive operating parameters. As an example of how machine learning can be implemented with ANI to save on human resources, particularly when undertaking mundane tasks, the Pentagon’s Algorithmic Warfare Cross-Functional Team, known as Project Maven, helps process surveillance data in order to prioritise potential targets for future offensive action, capable of doing so far quicker than even the most highly trained human analysts37. The system itself operates on a highly controlled and narrow platform, dependent on human control. This is the dominant direction in which western models of weaponised AI are moving. Such ANI technology is also being incorporated by the UK defence industry. The latest technology is currently being developed by the Defence Science and Technology Laboratory (DSTL) and the Ministry of Defence to develop a platform, known as SAPIENT, which can autonomously identify potential threats to soldiers on the battlefield, and send that data to the troops remotely38. This has the potential to reduce human error and relieve human resources required to monitor video footage. Now funded exclusively by DSTL, SAPIENT, standing for Sensors for Asset Protection using Integrated Electronic Network Technology, uses automation and ANI to ensure that the military user is presented with the information which they need at the time that they need it. This includes unusual or suspicious activity including people near a checkpoint or changes in behaviour. SAPIENT was trialled in 2018 in Canada as part of the Contested Urban Environment experiment (CUE 18); bringing together the Five Eyes allied nations to put the very latest ANI technology in the hands of soldiers on the ground. This included a range of unmanned aerial and ground vehicles, whilst technologies were also used to relay information to an operations centre for analysis by scientists and military personnel, in addition to planes sending autonomously refined information back to human operators below39. Combining all of these technologies from across the different nations, it was possible to generate information that could be fed to soldiers and military commanders – significantly enhancing their situational awareness.

The UK’s AI policy

A June 2018 Lord’s Select Committee set out its recommendations regarding the UK’s AI policy. The emphasis of the report was to develop an ethical approach to AI which would benefit UK and global society, whilst simultaneously ensuring that the UK remains a market leader in the field40. The UK must seek to actively shape AI’s development and utilisation, or risk passively acquiescing to its many likely consequences. By establishing an ethical AI policy, the UK can lead by example in the international community. The report recommends that the Government convene a global summit of governments, academia and industry to establish international norms for the design, development, regulation and deployment of artificial intelligence. Furthermore, the UK contains internationally leading AI companies, in addition to a dynamic academic and research culture, and so is in a position of strength to be among the world leaders in AI’s development. Given that UK law is used as a legal model for business the world over, this could lead to issues surrounding ethics and privacy in AI receiving closer attention and broader international agreement, hence further strengthening the UK’s position. Artificial intelligence, handled carefully, could be a great opportunity for the UK economy. However, the requirement to protect society from potential threats and risks remains paramount; central to this is an ethical approach, guided by shared principles.

What is the state of autonomy in weapon systems?

Within the broader discussion of RW, the rise of technology and autonomy was identified as a rising phenomenon back in 201741; in particular the need for addressing how autonomous weapons will change the landscape of military engagement. The rest of this paper seeks to address this concern.

In 2018 the Chief of the General Staff, General Mark Carleton-Smith, described how warfare is increasingly moving into non-traditional spaces, in areas such as cyber, artificial intelligence and, crucially, autonomous technology42. Central to developing ANI capability further, autonomous technology is at the heart of AI and robotics development. Autonomy has many definitions and interpretations but is generally understood to be the ability of a machine to perform an intended task without human intervention, using interaction of its sensors and

computer programming with the environment43. The feasibility of autonomy depends on (a) the ability of software developers to formulate an intended task in terms of a mathematical problem and a solution; and (b) the possibility of mapping or modelling the operating environment in advance. The use of machine learning in weapon systems is still experimental, as it continues to pose fundamental problems regarding predictability. Autonomy is already used to support various capabilities in weapon systems, including mobility, targeting, intelligence, interoperability and health management. Automated target recognition (ATR) systems, the technology that enables weapon systems to acquire targets autonomously, has existed since the 1970s. ATR systems still have limited perceptual and decision-making intelligence however. Their performance rapidly deteriorates as operating environments become more cluttered and weather conditions deteriorate, one of their current inherent weaknesses. This ‘complexity of context’ adds a challenge. Does a local national carrying an AK represent a target or just a farmer protecting himself? The software for the algorithms are still yet to fully develop to mitigate against this. This reinforces that the ability of humans to discern targets is still significantly greater than that of electronic processing algorithms44. Existing weapon systems that can acquire and engage targets autonomously are mostly defensive systems. These are operated under human supervision and are intended to fire autonomously only in situations where the time of engagement is deemed too short for humans to be able to respond. These have seen great success in current operations where wide area search analysis can be conducted by ATR systems much more effectively than by human control45.

Loitering weapons are the only ‘offensive’ type of weapon system that is known to be capable of acquiring and engaging targets autonomously. The loitering time and geographical areas of deployment, as well as the category of targets they can attack, are determined in advance by humans, who retain overall command and control, thereby restricting the degree of autonomy in these platforms. First developed by Israeli firm IAI in 2005, the Harop can cruise within a six-hour timeframe and deliver a precision strike 15 kg warhead. This technology has since been developed and sold by other states, including the US, China, Iran, Turkey and other smaller Eurasian states including Ukraine, Belarus and Azerbaijan. The US field the SwitchBlade, a drone killer missile system which has been utilised in both Afghanistan and Syria by SOF personnel46. The proliferation of both drone technology and autonomous platforms has led directly to a sharp increase in non-state actors utilising loitering weapons, including their use by IS.

What are the drivers of, and obstacles to, the development of autonomy in weapon systems?

The three main drivers to the continued development of autonomy in weapon systems are; strategic, operational, and economic. The US recently cited autonomy as a cornerstone of its strategic capability calculations, in addition to developing AI and robotics, in a bid to modernise defence plans in arenas it believes crucial to potential future military environments47. This seems to have triggered reactions from other major military powers, notably China who are keen to maintain certain geopolitical advantages over the US in the South China Sea. The second driver is operational. Military planners believe that autonomy enables weapon systems to achieve greater speed, accuracy, persistence, reach and coordination on the battlefield. In essence, aiding lethality in order to maintain combat effectiveness. Finally, the economic benefits are driving development also. Autonomy is believed to provide opportunities for reducing the operating costs of various weapon systems and weaponised UAV technology, specifically through a more efficient use of manpower.

However, there are also numerous challenges to developing autonomy in weapon systems which require careful consideration. The first is technological. Autonomous systems need to be more adaptive to operate safely and reliably in complex, dynamic and adversarial environments; new validation and verification procedures must be developed for systems that are adaptive or capable of learning. Secondly, institutional resistance. Military personnel often lack trust in the safety and reliability of autonomous systems; some military professionals see the development of certain autonomous capabilities as a direct threat to their professional ethos or incompatible with the operational paradigms that they are used to. This structural obstacle is being mitigated by two factors; top-down education in the military in the form of a high strategic imperative which developing AI, robotics and weapon autonomy requires in order for western militaries to remain competitive. This can be evidenced in both the 2018 US National Defense Strategy48 and in the 2018 UK Modernising Defence Programme, where in

the latter the Secretary of State for Defence states that the UK military is “pursuing modernisation in areas like artificial intelligence, machine-learning, man-machine teaming and automation to deliver the disruptive effects we need in this regard”49. Secondly, by incorporating the latest technology with the military on rigorous testing and training exercises it enables the end user to develop confidence in the equipment, thus removing this institutional distrust of autonomy.

A significant obstacle is presented by the legal ramification. International law includes a number of obligations which restrict the use of autonomous targeting capabilities. It also requires military command to maintain, in most circumstances, some form of human control or oversight over the weapon system’s behaviour. This restricts the utility of such platforms, though at this early stage whereby ANI is being utilised, this will assist in trust being developed between the public and the technology. Tying in with this are the normative considerations. There are increasing pressures from civil society against the use of autonomy for targeting decisions, which makes the development of autonomous weapon systems a potentially politically sensitive issue for militaries and governments. Lastly, there remain pertinent economic obstacles. There are limits to what costs can be afforded by national armed forces, particularly among western states in NATO, the majority of whom consistently fail to meet the mandated 2% of GDP spending on defence. In addition, the defence acquisition systems in most arms-producing countries remain ill-suited to the development of autonomy.

The Stockholm International Peace Research Institute’s report, Mapping the Development of Autonomy in Weapon Systems, highlighted some thought-provoking and practical recommendations for the continued development of autonomy in weapon systems50. A central theme in pioneering further autonomy is the notion of ‘autonomy in weapon systems’, rather than autonomous weapon systems. Shifting the focus away from ‘full’ autonomy and exploring instead how autonomy transforms human control, will likely lead to a more productive debate. Even further, allowing the scope of investigation to be opened beyond the issue of targeting and acquisition, to take into consideration the use of autonomy for collaborative operations and intelligence processing, will serve the military necessity of freeing up personnel to perform other tasks. Finally, there is a need to investigate the options for preventing the risk of weaponisation of civilian technologies by non-state actors. This is highlighted by the documented instances of IS, amongst other terrorist and criminal organisations, managing to achieve a tactical effect on the battlefield due to the ease and accessibility of civilian drone technology combined with the increased availability of autonomous platforms. This is an area in which NGOs could have great utility, assisting in shaping the ongoing debate surrounding AI and autonomy in RW further.

Future role of UGVs in remote warfare: vision and developments

The UK’s current vision for the future role of Unmanned Ground Vehicles (UGVs) originates from the British Army’s “Strike Brigade” concept, as outlined in the Strategic Defence Security Review 201551. This review proposed that British ground forces should be capable of self- deployment and self-sustainment at long distances, potentially global in scope. By 2025, the UK should be able to deploy “a war-fighting division optimised for high intensity combat operations”; indeed, “the division will draw on two armoured infantry brigades and two new Strike Brigades to deliver a deployed division of three brigades.” Both Strike Brigades should be able to operate simultaneously in different parts of the world, and by incorporating the next generation autonomous technology currently being trialled by the British Army, will remain combat effective post-Army 2020.

The ability for land forces of this size to self-sustain at long range places an increased demand on logistics and the resupply chain of the British Army, which has been shown to have been overburdened in recent conflicts. This is likely to increase due to the evolving nature of warfare and of the environments in which conflicts are likely to occur, specifically densely populated urban areas. These environments are likely to become more cluttered, congested and contested than ever before. Therefore, a more agile and flexible logistics and resupply system, able to conduct resupply in a more dynamic environment and over greater distances, is required to meet the challenges of warfare from the mid-2020s and beyond.

This will represent something of a shift in the UK’s vision for UGV technology, having previously been utilised almost exclusively for Explosive Ordnance Disposal (EOD) and Countering-Improvised Explosive Devices (C-IED) for both the military and the police, as opposed to being truly a force-multiplier developing the logistics and resupply chains. EOD and C-IED UGVs have been used by the UK since 1972 with the Wheelbarrow Mk 8B remote- controlled EOD robots, which are to be replaced by the US manufactured Harris T7 unmanned ground vehicle, due to enter service in 202052.

The MOD’s DSTL is developing this vision further, currently leading a three-year research and development programme entitled Autonomous Last Mile Resupply System (ALMRS)53. This research is being undertaken to demonstrate system solutions which aim to reduce the logistical burden on the Armed Forces, in addition to providing new operational capability and to reduce operational casualties. Drawing on both commercial technology as well as conceptual academic ideas – ranging from online delivery systems to unmanned vehicles – more than 140 organisations from small and medium-sized enterprises, to large military- industrial corporations, submitted entries.

The first phase of this programme challenged industry and academia to design pioneering technology to deliver vital supplies and support to soldiers on the front line, working with research teams across the UK and internationally. This highlights the current direction with which the British vision is orientated regarding UGVs, i.e., support-based roles. Meanwhile, the second phase of the ALMRS programme started in July 2018 and is due to last for approximately twelve months. It included ‘Autonomous Warrior’, the Army Warfighting Experiment 18 (AWE18), a 1 Armoured Infantry Brigade battlegroup-level live fire exercise, which took place on Salisbury Plain in November 2018. This saw each of the five remaining projects left in the ALMRS programme demonstrate their autonomous capabilities in combined exercises with the British Armed Forces, the end user. This provided DSTL with user feedback, crucial to enable subsequent development; identifying how the Army can exploit developments in robotics and autonomous systems technology through capability integration.

Among the final five projects short-listed for the second phase of ALMRS and AWE18 was a UGV multi-purpose platform called TITAN, developed by British military technology company QinetiQ, in partnership with MILREM Robotics, an Estonian military technology company. Developing its Tracked Hybrid Modular Infantry System (THeMIS), the QinetiQ-led programme impressed in the AWE18.

The THeMIS platform is designed to provide support for dismounted troops by serving as a transport platform, a remote weapon station, an Improvised Explosive Device (IED) detection and disposal unit, and surveillance and targeting acquisition system designed to enhance a battlefield commander’s situational awareness. THeMIS is an open architecture platform, with subsequent models based around a specific purpose or operational capability.

THeMIS Transport is designed to manoeuvre equipment around the battlefield to lighten the burden of soldiers, with a maximum payload weight of 750 kilograms. This would be adequate to resupply a platoon’s worth of ammunition, water, rations and medical supplies and to sustain it at 200% operating capacity – in essence, two resupplies in one. In addition, when utilised in battery mode it is near-silent and can travel for up to ninety minutes. When operating on the front-line, this proves far more effective than a quad bike and trailer, which are presently in use with the British Army to achieve the same effect. This is often overseen by the Platoon Sergeant, the platoon’s Senior Non-Commissioned Officer and most experienced soldier. Relieving this individual of such a burden would create an additional force multiplier during land operations.

THeMIS can also be fitted to act as a Remote Weapons System (RWS), with the ADDER version equipped with a .51 calibre Heavy Machine Gun (HMG), outfitted with both day and night optics. Additional THeMIS models include the PROTECTOR RWS, which integrates Javelin anti-tank missile capability. Meanwhile, more conventional THeMIS models include GroundEye, an EOD UGV, and the ELIX-XL and KK-4 LE, which are surveillance platforms that allow for the incorporation of remote drone technology.

How UGVs will influence the capabilities and tactics of small infantry units in a remote warfare capability

In order to comprehend how UGVs will influence the capabilities and tactics of small infantry units, it is important to first understand the different types of infantry units. There are currently 33 regular British Army infantry battalions, roughly comprising 400-500 personnel each, with each one specialising in one of the following roles; light-role, mechanised or armoured warfare. The break-down of these are; 22 light-role battalions, six armoured battalions, and five mechanised battalions54. Light-role battalions primarily do not have access to vehicles, though this often changes on operations, and are trained to move and fight primarily on foot. Mechanised battalions have access to vehicle-born weapons platforms, either Jackal 2 or Foxhound, which provide a moderate troop transporting capacity. Armoured battalions have access to armoured personnel carriers, the British Warrior, heavily armed and armoured and capable of delivering the infantry dismounts inside into the heart of modern battle, ensuring that they arrive safely and have a strong fire-support base from which to launch their attack. It is worth noting that in both the mechanised and armoured roles, infantry battalions are relatively self-sufficient in the manner of both direct and indirect fire support, and of logistics and resupply, the two areas where UGVs have the most utility for an infantry battalion.

There is far greater potential for the development of UGVs within a light-role infantry battalion, due to the very fact that the majority of what organic fire support and resupply they have available to them is man-portable and so can limit the effect that they wish to achieve, unless reliant on external units for support, such as transport. Therefore, this analysis will seek to determine how UGVs can influence the capabilities and tactics of light-role infantry units. In analysing UGV influence on light-role infantry battalions’ tactics, it is necessary to determine what their roles on the battlefield are. Infantry battalions have two overall roles in battle; offensive, and defensive. These can be further broken down into more mission-specific roles, as required, though for the most part they do fall into these two operational capabilities. By analysing each in turn, it is possible to further elucidate how UGVs can influence the capabilities and tactics of light-role infantry battalions.

In an offensive environment, UGV capability can be utilised in two dominant roles. First will be the further consideration of RWS and direct fire-support, and second in a support role within the transport and resupply chain. Whilst there remains much room for discussion surrounding the ethical implications and employment of RWS in today’s conflicts, there is certainly a requirement for this capability, and it comes down almost exclusively to manning. During a deliberate attack on a fortified enemy position, an RWS platform, such as the THeMIS ADDER with a Heavy Machine Gun (HMG), can provide the direct fire-support which would take either a mechanised platform such as Jackal 2 or MASTIFF, or a 2 – 3-man HMG crew, with accompanying quad bike and trailer. In a light-role configuration, clearly the mechanised platform is inappropriate, leaving the only viable option for such a battlefield effect requiring a 3-man team plus vehicle. This situation is replicated with other light-role infantry weapon systems, including the 40mm Grenade Machine Gun (GMG) and even the 81mm mortar. If UGV technology can be utilised to fulfil these additional direct and indirect fire roles, then this would significantly free up manpower. There is a further advantage of utilising UGVs in this role. During an offensive operation, whenever the battle moves forward, then so does the fire support. If this is coming from a static 3-man fire support team, then the team will have to cease activity whilst dismantling the weapon system, moving forwards to a new position, and setting it back up again. This is not only time consuming but will leave large periods where there is no fire support being provided and thus leaving the ground troops vulnerable. Clearly, UGV platforms in this capacity, from a purely operational perspective, would provide great utility and cut down considerably on manning requirements, freeing up other ground troops to other parts of the battlefield.

The transport and resupply chain can also be enhanced during offensive operations. During this phase of operations, there are often many offensive actions coordinated simultaneously, stretching the resupply chain to its limit. Again, this comes down to manning and equipment; often there are simply too few personnel available, with limited transport capacity, to facilitate a light-role infantry battalions’ requirements. This can result in a variety of scenarios ranging from time wastage to sub-unit combat ineffectiveness. The THeMIS transport platform can transport small numbers of personnel around the battlefield, at a much-increased pace than by foot, serving several key objectives. This is ideal for extracting casualties back to receive treatment, which in battle is conducted by a four-man team per casualty, and then by a battlefield ambulance operated by a minimum of two personnel. In an incident involving two casualties requiring extraction ten personnel will be utilised. Removing personnel from the battle is a labour-intensive, often during critical moments. A THeMIS transport platform can be set remotely for pre-designated waypoints, can carry at least two casualties, and at a much- increased speed of extraction than by foot. Not only does this potentially result in a casualty receiving much quicker treatment, it additionally frees up vital manpower during the battle.

In defensive operations, Infantry battalions may be required to hold key terrain for a prolonged period of time. This has a limiting effect on other capabilities which commanders may seek concurrently. For instance, the ability to hold a piece of key terrain or ground would require a surveillance capability in addition to the ability to defend the ground through weapons considerations. In order to achieve this, these tasks require manpower which cannot be used in other battlefield roles whilst conducting defence, such as patrols conducted by sub-units on foot. Therefore, UGV technology could be utilised as a surveillance asset at certain pieces of key terrain, in addition to a RWS to guard that piece of terrain against enemy attacks, with large-scale area denial achieved by emplacement of interlocking weapon systems (THeMIS ADDER with mounted GMG and HMG with ranges of 2 kilometres plus). The combination of a THeMIS ELIX-XL with an ADDER RWS could provide this capability and thus reduce manpower at critical points on the ground whilst conducting defensive operations, the ELIX- XL drone having the capability to be autonomously launched and recovered by the UGV Drone Nest55. This ability also reduces the need to launch standing patrols between the enemy’s defences and friendly forces defences; a key Infantry task whilst operating in defence. The ELIX-XL surveillance platform incorporates a real time video from two on-board cameras and an on-board video recorder. In addition, it provides a fully autonomous flight control system and an autonomous mission execution control system, with the ability to select active waypoints and change the coordinates in flight.

  Weather and optics dependent, this battlefield asset would be able to achieve highly comparable results with a small reconnaissance patrol. These patrols, encompassing eight individuals, are often launched in sequence, comprising three separate patrols at a time. The ELIX-XL has the ability to switch drone batteries, minimising the time between flights. This surveillance capability would therefore be able to theoretically conduct the same workload as a platoon of Infantry (30 men) whilst operating standing patrols during a defensive phase of operations, further freeing up manpower across the Infantry battalion to conduct other mission- specific tasks.

The main challenges with introducing UGVs in combat functions

If Britain is to maintain the momentum which it has so far developed in pioneering the utility of UGVs, it should take seriously the following challenges. Moreover, it should see each one as a possibility for strengthening its global role in this field by seeking to convert these challenges into opportunities for constant refinement. The main challenges introducing UGVs further within the military can be broadly based around four key areas; operational, institutional, financial and legal. Whilst the challenges surrounding these range in difficulty to address, each provide potential for doing so.

Whilst the proposed merits of utilising UGVs in combat functions by Infantry units have been discussed above, it is only through extensive operational trials of UGVs within ground close combat units that a more accurate and reliable assessment can be made regarding the fulfilment of this hopeful utility. Seeking to build on the success of the Royal Navy-led Unmanned Warrior exercise in 2016, the DTSL and British Army-led Army Warfighting Experiment, Armoured Warrior, sought to push the existing boundaries of technology and military capability in the land environment during the extensive month-long exercise with 1 Armoured Infantry Brigade in November 2018. This allowed time for the operationalising of the various capabilities of the UGV being trialled. Crucially, as previously discussed, this enabled the technology to be run by the desired end-user, the British Infantry, and provided critical operational information allowing for further integration of UGVs further into Army 2020 and beyond. Armoured Warrior assisted significantly in reducing the operational challenges faced by the military, which require timing and resources in order to mitigate complex operational compatibility with the end-user.It can often take time for policy to transition from the strategic space to the operational space, and institutional barriers can elongate this process further still. However, in his first major speech as the Chief of the General Staff, General Mark Carleton-Smith stated how “the nature of warfare is broadening beyond the traditional physical domains”56, adding that the 21st Century battlefield requires non-traditional skills in an effort to retain lethality. This clearly demonstrates how the nature of warfare and its constant evolution is being understood by the highest military levels, offsetting the potential for institutional delays in bringing the latest technology to the battlefield.

Moreover, General Carleton-Smith stated the need to place “some big bets on those technologies that we judge may offer exponential advantage because given the pace of the race, to fall behind today is to cede an almost unquantifiable advantage from which it might be impossible to recover.”57 The challenge therefore is the need to implement this strategic outlook right down the military chain, to the end-user; Infantry battalion commanders. Through continued integration of UGVs into brigade-level and below training exercises, the exposure this will give to these sub-unit commanders will significantly reduce the institutional gaps between strategic doctrine and operational delivery.

UGV technology is expensive to fund, though there are multiple avenues for research and development. On 16 September 2016 the Defence Secretary launched the Defence Innovation Initiative (DII)58. This is an £800 million fund designed to increase the pace of development of various defence projects, aimed at the private sector in addition to academia and research institutions. Autonomous Warrior and the UGV technology pioneered through the programme is funded by the DII and seen as integral to the future development of both British defence and industry. In addition, the 2018 Modernising Defence Programme set aside £160 million for the Defence Transformation Fund to invest in new concepts and technology aimed at improving autonomous capability59. This figure is expected to rise to £500 million as part of the 2019 Spending Review.

The discussion on legality, ethics and meaningful human control with the reality of weapon systems development and weapon use needs further exploration and is almost certainly the weakest theoretical and subsequently operational component of further UGV integration into the military. To begin with, refocussing of the legality discussion to one of the development of ‘autonomy in weapon systems’ rather than autonomous weapons or LAWS as a general category, would enable the debate to progress rather than regress. In addition, by shifting the focus away from ‘full’ autonomy and exploring instead how autonomy transforms human control, greater engagement will be seen from various stakeholders likely to raise issues of legality regarding full autonomy. Furthermore, every effort should be made by the UK government to engage with the Convention on Certain Conventional Weapons (CCW) regarding the development of UGV technology and the legal implications of autonomy in weapon systems on the battlefield. By seeking to actively engage with the CCW on these matters, Britain can ensure that it maintains its global role further in the continued development of UGV and RWS technology, in their subsequent implementation and operationalisation.

Concluding remarks

By seeking to understand further the roles within the British military both AI and robotics currently have, in addition to what drives these roles and what challenges them, it is possible to gauge the continued evolution of RW with the emergence of such technologies. Specifically, UGVs and RWS’ which were trialled extensively in 2018 by the British Army. Based around research conducted on these recent trials, combined with current up-to-date in-theatre applications of such technology, it is assessed that the use of such equipment will expediate the rise of RW as the preferred method of war by western policy makers in future low to medium level intensity conflicts seeking to minimise the physical risks to military personnel in addition to engaging in conflict more financially viable.

Acknowledgments

I am gratitude for advice, comments, and discussions with Vivian Barlow, Kyle Orton, and Rob Spalton. I am also grateful to those who contributed in some manner, but who must remain unnamed. All mistakes remain exclusively the authors.

1 https://shape.nato.int/resources/site631/saceur/documents/daalder_stavridis_final.pdf

2 https://www.tandfonline.com/doi/pdf/10.1080/01402390.2015.1045494?needAccess=true

33 https://www.parliament.uk/business/publications/written-questions-answers-statements/written- question/Commons/2018-10-29/184962

4 https://www.theguardian.com/world/2015/oct/14/obama-deployment-us-troops-cameroon-boko-haram

5 https://www.oxfordresearchgroup.org.uk/Handlers/Download.ashx?IDMF=0232e573-f6d6-455e-9d34- 0436925002d4

6 https://publications.parliament.uk/pa/cm201314/cmselect/cmdfence/197/197vw06.htm

7 https://publications.parliament.uk/pa/cm201314/cmselect/cmdfence/197/197vw06.htm

8 https://www.theguardian.com/politics/2019/feb/11/uk-will-deploy-drone-squadrons-after-brexit-says- defence-secretary-gavin-williamson

99 https://www.gov.uk/government/news/new-medal-awarded-to-recognise-the-changing-character-of- warfare

10 https://www.cfr.org/timeline/iraq-war

11 http://iraqslogger.powweb.com/downloads/IraqReport9.pdf

12 https://publications.parliament.uk/pa/cm200708/cmselect/cmdfence/982/98206.htm 13 https://www.brookings.edu/wp-content/uploads/2016/06/singer_20020701.pdf

14 https://www.brookings.edu/wp-content/uploads/2016/06/singer_20020701.pdf

15 https://www.govinfo.gov/content/pkg/PLAW-109publ163/pdf/PLAW-109publ163.pdf 16 https://www.gao.gov/assets/100/94660.pdf p.1.

17 https://www.gao.gov/assets/700/692152.pdf p.2.

18 https://media.defense.gov/2017/Dec/19/2001858653/-1/-1/1/DODIG-2017-099.PDF.

19 https://www.thenation.com/article/american-special-ops-forces-have-deployed-to-70-percent-of-the- worlds-countries-in-2017/

20 https://www.gao.gov/assets/680/671462.pdf

21 https://www.gao.gov/assets/680/671462.pdf

22 https://foreignpolicy.com/2017/08/01/donald-trump-is-pushing-americas-special-forces-past-the-breaking- point-jsoc-navy-seal/

23 https://www.cfr.org/blog/us-military-presence-sub-saharan-africa-will-likely-decline

24 https://apps.dtic.mil/dtic/tr/fulltext/u2/1026959.pdf p.10.

25 https://www.washingtonpost.com/news/checkpoint/wp/2017/12/06/there-are-four-times-as-many-u-s- troops-in-syria-as-previously-acknowledged-by-the-pentagon/?utm_term=.15b7eff85c69

26 www.rudaw.net/english/middleeast/syria/03012018

27 https://www.nytimes.com/2019/01/18/us/syria-soldiers-killed.html

28 https://apps.dtic.mil/dtic/tr/fulltext/u2/1017816.pdf p.22. 29www.understandingwar.org/sites/default/files/The%20Pitfalls%20of%20Relying%20on%20Kurdish%20Force s%20to%20Counter%20ISIS_0.pdf

30

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/765879/ ModernisingDefenceProgramme_report_2018_FINAL.pdf

31 https://www.c4isrnet.com/it-networks/2018/07/17/heres-how-much-a-new-artificial-intelligence-center- could-cost/

32 https://admin.govexec.com/media/establishment_of_the_joint_artificial_intelligence_center_osd008412- 18_r….pdf p.2.

33 https://www.roboticsbusinessreview.com/manufacturing/global-spending-on-robots-drones-to-top-115b- in-2019-says-idc/

34 https://internetofbusiness.com/pentagon-military-self-driving-vehicles/ 11

35 https://www.techworld.com/startups/deepmind-co-founder-says-general-ai-remains-long-way-off- 3651343/

36 https://philpapers.org/archive/MLLFPI.pdf

37 https://blogs.nvidia.com/blog/2017/11/01/gtc-dc-project-maven-jack-shanahan/

38 https://www.gov.uk/government/news/streets-ahead-british-ai-eyes-scan-future-frontline-in-multinational- urban-experiment

39 https://www.gov.uk/government/news/streets-ahead-british-ai-eyes-scan-future-frontline-in-multinational- urban-experiment

40 https://publications.parliament.uk/pa/ld201719/ldselect/ldai/100/10002.htm

41 https://www.oxfordresearchgroup.org.uk/Handlers/Download.ashx?IDMF=0232e573-f6d6-455e-9d34- 0436925002d4 p.2.

42 https://rusi.org/annual-conference/rusi-land-warfare-conference/cgs-keynote-address-2018

43 https://www.sipri.org/sites/default/files/2017- 11/siprireport_mapping_the_development_of_autonomy_in_weapon_systems_1117_1.pdf p.18. 44file:///C:/Users/Owner/AppData/Local/Packages/Microsoft.MicrosoftEdge_8wekyb3d8bbwe/TempState/Do wnloads/072001_1%20(1).pdf p.7 45file:///C:/Users/Owner/AppData/Local/Packages/Microsoft.MicrosoftEdge_8wekyb3d8bbwe/TempState/Do wnloads/072001_1%20(1).pdf

46 https://www.businessinsider.com/us-special-ops-just-got-350-kamikaze-drones-to-fight-isis-in-iraq-2017- 5?r=US&IR=T

47 https://dod.defense.gov/Portals/1/Documents/pubs/2018-National-Defense-Strategy-Summary.pdf p.7. 48 https://dod.defense.gov/Portals/1/Documents/pubs/2018-National-Defense-Strategy-Summary.pdf

49https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/765879 /ModernisingDefenceProgramme_report_2018_FINAL.pdf p.15.

50 https://www.sipri.org/sites/default/files/2017- 11/siprireport_mapping_the_development_of_autonomy_in_weapon_systems_1117_1.pdf p.118.

51https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/478933 /52309_Cm_9161_NSS_SD_Review_web_only.pdf

52 https://www.gov.uk/government/news/british-army-receives-pioneering-bomb-disposal-robots

53 https://www.gov.uk/government/publications/accelerator-competition-autonomous-last-mile- supply/accelerator-competition-autonomous-last-mile-resupply

54 https://www.army.mod.uk/who-we-are/corps-regiments-and-units/infantry/

55 https://milremrobotics.com/product/themis-elix-xl/

56 https://defpost.com/british-army-launches-autonomous-warrior-land-experiment-rusi-land-warfare- conference/

57 https://defpost.com/british-army-launches-autonomous-warrior-land-experiment-rusi-land-warfare- conference/

58 https://www.gov.uk/government/publications/accelerator-themed-competition-revolutionise-the-human- information-relationship-for-defence/defence-innovation-initiative-information

59

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/765879/ ModernisingDefenceProgramme_report_2018_FINAL.pdf

References

Aoi, C. (2011). Legitimacy and the Use of Armed Force. London: Routledge.

Bankoff, G. (2003). Regions of Risk: Western Discourses on Terrorism and the Significance

of Islam. Studies in Conflict & Terrorism. 26(6):413-428.

Blanchard, C.M. and Humud, C.E. (2017). The Islamic State and US Policy. Congressional

Research Service Washington United States.

Boeke, S. and Schuurman, B. (2015). Operation ‘Serval’: A strategic analysis of the French

intervention in Mali, 2013–2014. Journal of strategic studies, 38(6). pp.801-825.

Byman, D. and Merritt, I.A. (2018). The New American Way of War: Special Operations Forces

in the War on Terrorism. The Washington Quarterly, 41(2). pp.79-93.

Daalder, I.H. and Stavridis, J.G. (2012). NATO’s victory in Libya: the right way to run an

intervention. Foreign Affairs. pp.2-7.

Federici, V., (2015). The rise of Rojava: Kurdish autonomy in the Syrian conflict. SAIS Review

of International Affairs, 35(2). pp.81-90.

Gross, O. (2015). The new way of war: is there a duty to use drones. Fla. L. Rev., 67, p.1.

Müller, V.C. and Bostrom, N. (2016). Future progress in artificial intelligence: A survey of expert opinion. Fundamental issues of artificial intelligence. Springer, Cham. pp. 555-572.

Ratches, M. (2011). Review of current aided/automatic target acquisition technology for military target acquisition tasks. Optical Engineering 50(7).

Smith, K.W. (2015). Drone technology: Benefits, risks, and legal considerations. Seattle J. Envtl. L., 5, p.i.

Watts, T., and Biegon, R. (2017). Defining Remote Warfare: Security Cooperation. Oxford Research Group. London.

25

About the author

Rob Clark

Robert clark is a post-graduate researcher at Kings Collage London, and a British Army veteran. His expertise is British Defence engagement and Indo-Pacific Security.

Related posts

Military Learning in the 21st Century

The Wavell Room Team

Computational Advantage; Defence’s Missed Opportunities in Artificial Intelligence

Peter A

An interview with Major-General Stephen Cadden

Major-General Cadden