BY RICHARD NGHIEM

UNCREWED SYSTEMS

Over the last decade, the number of countries investing in Artificial Intelligence (AI), specifically for military purposes, has rapidly increased as it is now widely believed that the world is on the brink of another military revolution. In the eyes of many, lethal autonomous weapon systems (LAWS) will soon become the most pervasive and decisive form of combat power on the battlefield that will redefine how war is waged in the 21st century. They will guarantee credible force projection in highly contested anti-access area denial (A2/AD) environments, drastically increase the speed of “kill chains”, and take on a much more significant decision making role at the tactical level.

These claims, while bold, have arguably already been substantiated and validated by the current war in Ukraine in which makeshift factories and labs have been constructed all across the country to manufacture AI-enabled autonomous weapon systems such as armed uncrewed ground vehicles, kamikaze sea drones, and cheap loitering munitions. In their fight against the Russian Army, outnumbered Ukrainian troops have had to extensively rely on autonomous drones and weapon systems as they are more survivable and resistant against electronic jamming. With Uncrewed Aerial Vehicles (UAVs), Uncrewed Ground Vehicles (UGVs), and Uncrewed Surface Vehicles (USVs) operating by themselves, operators can simply lock onto a target and let the combat platform do the rest. Furthermore, Ukrainian engineers, deployed in the field, have also built hundreds of UAVs based on AI models which have been trained with large amounts of data from real life frontline drone missions. This meticulous work has enabled Ukrainian soldiers to individually operate up to seven drones on bombing and reconnaissance missions deep inside Russian-held territory. Lastly, UGVs, equipped with gun turrets with autonomous targeting that uses AI-trained software have also been seen in combat, automatically tracking and shooting at Russian troops.

Consequently, having witnessed the capabilities of AI-enabled autonomous weapon systems on the battlefield in Ukraine for the past two years, countries around the world are now scrambling to gain dominance in this technological area and are actively seeking to harness the power of AI for military advantage. For example, the Canadian Armed Forces (CAF) and the Department of National Defence (DND) released their ‘Artificial Intelligence Strategy’ in March of this year. As another example, China has made clear its ambition to become the world leader in AI by 2030 while President Vladimir Putin has highlighted AI research as a top priority for Russia as he believes that “whoever becomes the leader in this sphere will become the ruler of the world.” The US, alarmed by the challenge posed by both of these countries, has also entered the technological race, committing to a “third offset” strategy in which it will invest heavily in AI, autonomy, and robotics to sustain its military superiority. The side with the fastest, most effective ‘kill chain’ will win in a modern war.

From an academic perspective, the kill chain concept is the ability of a military to rapidly and accurately identify, fix, track, target, and engage an enemy target.

LEVERAGING AI FOR THE CAF

Given the growing importance of faster and more efficient ‘kill chains’ and the start of an autonomous weapon systems arms race, Canada should also look to exploit the power of AI for its own armed forces. Given the defence and security challenges posed by Canada’s vast geographic size and limited military manpower, AI-enabled uncrewed combat platforms, in the form of long endurance UAVs, Uncrewed Combat Aerial Vehicles (UCAVs), Uncrewed Undersea Vehicles (UUVs), and USVs, could prove to be of enormous value to the CAF. These assets, first of all, are currently able to stay in the air and on or under the surface of the country’s oceans longer than most of the military’s planes, ships, and submarines. Consequently, they are highly likely to be extremely effective at providing year-round surveillance and ensuring control of the entirety of Canadian airspace, coastlines, and maritime approaches. Second, the cost of patrol missions carried out by autonomous drones over large swaths of territory is only a fraction of that of crewed platforms. Lastly, these uncrewed combat systems would exponentially enhance the CAF’s situational awareness, accelerate its ability to quickly close ‘kill chains’ to speeds unattainable by even the most well-trained servicemen and servicewomen, and enable Canadian commanders to be considerably more aggressive in using them for combat operations as there is zero risk to CAF personnel.

CAF PERSONNEL SHORTAGE

This reduced reliance on human personnel is particularly intriguing for the Canadian military as it is currently facing a significant recruiting and retention crisis. Defence Minister Bill Blair has already estimated that the military is short up to 16,500 members and said the Armed Forces’ failure to boost recruitment is leading it into a “death spiral.” The hole in which the DND finds itself now, however, may be deeper and potentially more serious than it might appear at first glance, in part because of all the new equipment the federal government has ordered, or plans to order in the near future as part of its new defence policy, “Our North, Strong and Free,” including an expanded submarine fleet, F-35 fighters, MQ-9B SkuGuardian drones and long-range ground-based missiles. To properly operate these combat assets, carry out the new defence policy, and fulfill the demands of modernizing continental defence under NORAD, former top military commander, General Wayne Eyre, has stated that the DND would need to take on an additional 14,500 people on top of the 16,500 required to bring the military up to authorized strength. Reserve forces, the CAF’s insurance policy, would also have to reach 30,000 personnel. Unfortunately, though, with the armed forces facing a relevancy problem along with diminishing awareness and interest among young people, current DND projections do not foresee the military having any chance of returning to its current authorized strength of 71,000 regular and 30,000 reserve forces. As a result, the CAF must look at other solutions including autonomous uncrewed combat systems as they could potentially be the only answer to its personnel shortage problem.

MARITIME DOMAIN

For example, given the RCN’s recent struggles to grow its force size to fulfill the anticipated needs of the Canadian Surface Combatant (CSC) and other new platforms, employment of Large Uncrewed Surface Vehicles (LUSVs) and Medium Uncrewed Surface Vehicles (MUSVs) by the service branch would help reduce personnel requirements and greatly augment the capabilities and capacity of the current fleet. From LUSVs supporting offensive operations across Air, Surface, and Underwater warfare to MUSVs supporting Intelligence Surveillance & Reconnaissance (ISR), Mine Countermeasures (MCM) and Electronic Warfare (EW) operations, the Navy could operate a large fleet of autonomous maritime drones, increasing its overall size by 70% and adopting a more distributive approach in terms of maritime operations and capabilities in a fiscally responsible manner.

These new assets would also alter the nature of how the RCN operates by radically increasing the operational availability of its vessels to respond to ever increasing demands from Ottawa. This is due to the fact that USVs require less maintenance, are highly responsive and adaptable, and are ready to deploy at a moment’s notice. Most importantly, the interoperable nature of combat capable maritime drones with traditionally crewed vessels and their capacity to sail independently, would allow the RCN to project significantly more influence and presence around the world at reduced operating costs.

While increased combat capacity and power projection capabilities, generated by the use of uncrewed assets, would unquestionably be welcomed by Ottawa, there are still some defence analysts who remain doubtful of the operational viability of USVs or UUVs in actual wartime. To allay such concerns, one only needs to look at the successes the US Navy has achieved with its AI-enabled vessels. These new ships have demonstrated that they not only possess full command and control along with growing war fighting capabilities but can also safely navigate from one location to another without any collisions or need for human intervention.

For example, the US Navy’s Uncrewed Surface Vessel Division One’s (USVDIV-1) USVs Ranger, Mariner, Seahawk and Sea Hunter successfully arrived in Sydney, Australia, completely autonomously, for a scheduled port visit on October 24, last year, ahead of their participation in bilateral exercises with the Royal Australian Navy. Civilian crews closely monitored these maritime drones’ journeys, but were under strict orders to not interfere unless absolutely necessary.

In addition to their navigational prowess, these naval assets have also showcased their growing war fighting capabilities with the outfitting of certain USVs with next-generation command and control systems and Aegis Combat Systems, and the test-firing of an SM-6 missile from an onboard MK 41 VLS canister in 2021. The USV Ranger launched the missile, successfully destroying an undisclosed target, once again reaffirming that autonomous maritime drones are indeed combat capable in a conventional war setting. Consequently, the procurement of a large number of these autonomous MUSVs, equipped with small numbers of VLS cells in standard shipping containers, should be seriously considered by Ottawa as it could be a comparatively inexpensive fix for the RCN’s current shortage in offensive missile firepower caused by delays in the CSC program.

Besides prioritizing autonomous MUSVs, missiles, command and control systems, and combat systems, the RCN should also rigorously study the benefits of “crewed-uncrewed teaming,” particularly NAVCENT’s Task Force 59’s project which demonstrated the ability of uncrewed autonomous platforms to pair with traditionally crewed ships to identify and target hostile forces at sea and then, using live loitering munitions launched from another uncrewed platform, successfully engage and destroy these maritime targets. Another example of this concept was the Sea Hunter and Sea Hawk teaming up together with a guided missile destroyer as part of RIMPAC 2022. During this exercise, USV Sea Hawk hauled a towed array sonar for Anti-Submarine Warfare (ASW) missions, contributing acoustic sensor data regarding potential undersea contacts directly to sonar operators aboard the USS P. Lawrence while the Sea Hunter autonomously operated an electronic warfare suite in collaboration with the destroyer to jam the radars of over-the-horizon ‘enemy’ ships. The Mariner and Ranger also joined in, serving as forward sensors for the fleet, to improve both detection and missile engagement capabilities.

However, while all these USVs, designed for anti-ship warfare, EW, ISR, and air defence, would be incredibly vital additions to Canada’s surface fleet, Ottawa also desperately needs to acquire a network of autonomous uncrewed surface and subsurface platforms that can operate as an unattended system in the Far North year-round to constantly search for and monitor Russian submarines and, if necessary, conduct undersea warfare. This is due to the fact that despite Russia’s poor military performance in its war against Ukraine, Moscow’s undersea capabilities remain among the best in the world, and as such, still represent a potent threat to Canadian and NATO security, particularly in the Arctic.

Having said that, Western navies spend an inordinate amount of resources for ASW and need to find a new, more efficient way of defending against submarines. The large-scale use of USVs in conjunction with UUVs could be the answer as the idea of deploying these assets to conduct persistent undersea surveillance, with high accuracy and without fatigue, and yet at a fraction of the cost of crewed platforms is extremely attractive. Ship-deployed UUVs along with ocean floor acoustic sensors and USVs, equipped with sonar towed arrays, would communicate and collaborate with each other to detect the heading, speed, and position of hostile targets while UAVs such the RCAF’s already procured SkyGuardian drones could also join the hunt with air-dropped sonobuoys. Once the enemy sub or subs have been detected and located, USVs would act as a relay / router and direct friendly attack submarines along with torpedo-armed UUVs and UAVs to conduct ASW operations against them, closing the ‘kill chain’.

After listening to such a scenario, many would likely find this operational concept to be highly tempting. Current Class III Medium Displacement Unmanned Surface Vehicle (MDUSVs) such as the US Navy’s Sea Hunter, an uncrewed self-piloted trimaran, only cost $15,000-$20,000 a day to operate compared with $700,000 a day for a major surface combatant like the CSC. They also have an impressive range of 10,000 nautical miles and can autonomously endure up to 90 days at sea without a crew. Given these facts, in the event of a war, a massive flotilla of cheap Sea Hunter USVs, instead of a few expensive destroyers, could more effectively search and scout a wider area for both hostile subs and mines in advance of Canada’s crewed naval task forces, detect incoming threats, and successfully neutralize them with anti-submarine torpedoes or mine countermeasures. Such a significant change in operational doctrine would greatly reduce the need for larger crewed warships as well as personnel, and alternatively prioritize the procurement of vessels like the Sea Hunter, transforming Canada’s ASW operations forever.

In addition to USVs like the Sea Hunter, uncrewed submersibles such as America’s new Manta Ray and Orca drones could also represent the future of undersea warfare. Unlike most UUVs which must be tethered to crewed ships for power and other technical support, both the Manta Ray and Orca have been designed to autonomously draw power from their operating environment to achieve almost unlimited range. The Manta Ray, for example, specializes in wave-energy conversion, or deriving electrical energy from the power of ocean and river currents. It leverages Reversed Electro Dialysis to generate a salinity gradient difference which transports ions that result in an electric potential, and therefore electricity. The undersea vehicle can also deploy a separate “Thermal Energy Pod” that creates electrical energy by harnessing the power of the ocean’s “thermal gradient,” or the mixing of warm and cooler currents.

Besides wave-energy conversion, in terms of autonomous navigation, when the Manta Ray needs to go up or down, it changes buoyancy for only a few minutes by pumping sea water to change the weight of the vehicle. As a result, the Manta Ray is able to take advantage of its unique, hydro-dynamically optimized hull to glide forward, ascend and descend, turn, and hover and anchor itself to the seafloor, using minimal power and energy, and “hibernate until needed” to perform whatever missions it might be configured for without the need for on-site human logistics or a human operator. Some of these missions include conducting ISR, mine-laying, and serving as a mothership for smaller UUVs.

Having said that, while Manta Ray-like drones should unquestionably be procured by the RCN, this platform is far from the only UUV that Ottawa should be interested in. Possessing the ability to evade detection by both passive and active sonar, the US Navy’s Orca vehicle is another undersea drone that the DND needs to consider. It is an uncrewed diesel-electric submarine, equipped with lithium-ion batteries and a shrouded propeller, capable of operating autonomously over thousands of miles and performing various missions with its 34-foot modular payload bay. Such combat tasks include underwater mining in highly contested waters, electronic warfare, intelligence gathering, long range strikes, and most importantly, undersea surveillance. Consequently, these eXtra Large Uncrewed Undersea Vehicles (XLUUVs), if purchased, will become an integral component of Canada’s subsurface fleet, particularly in the Arctic, either by shaping the battle-space in favour of the RCN by laying sea mines to vex enemy ships, or through being armed with torpedoes in preparation for full-scale ASW.

Lastly, the RCN requires an autonomous uncrewed platform that is capable of not only patrolling Canada’s vast coastlines and monitoring ocean approaches continuously, but also providing harbour security and replacing the fleet’s current crewed Kingston-class coastal patrol ships. This would free up a significant number of sailors who could be retrained and reassigned to the fleet’s future personnel-heavy CSCs without sacrificing coastal defence. Such a repurposing of talent is also crucial for being able to properly operate and maintain the Nation’s recently announced new fleet of 12 under-ice attack submarines.

Given this urgent need for an uncrewed option for coastal defence, a prime candidate is the Common Uncrewed Surface Vehicle (CUSV) from Textron Systems. From its specialization in minesweeping with Raytheon’s AN/AQS-20 and Northrop Grumman’s AN/AQS-24 sonars to ISR to harbour security, this USV is a mature, multi-mission, and multi-payload capable vehicle with significant operational experience. It can independently operate for 20+ hours or up to 1,400 nautical miles and is capable of a variety of missions including surface warfare with a remotely operated 0.50-caliber machine gun and two vertical-launch Hellfire missiles, EW, and signals intelligence. Finally, the CUSV is highly deployable as it is small enough to be launched from any ship with a well deck, the CSC’s Mission Bay Handling System, or from any regular pier in a harbour.

READ: UNCREWED SYSTEMS REPORT

AIR DOMAIN

This autonomous combat potential, however, is not limited to just the Navy. In the context of the air domain, the RCAF should adopt a CONOPS in which its newly procured crewed F-35s are paired with 4-5 ‘Loyal Wingman’ aircraft, each equipped with different payloads and responsible for different missions. These semi-autonomous AI-enabled UAVs could specialize in one of many combat effects such as electronic jamming, precision strike, Suppression of Enemy Air Defenses (SEAD), or ISR. This emphasis on large numbers of lower cost drones is absolutely key for an air force like Canada’s. Collectively, these autonomous systems will enable the RCAF to credibly operate within an enemy’s A2/AD threat envelope at a lower weapon-to-target cost-ratio and destroy extremely high value, expensive strategic targets.

If the RCAF is to remain operationally relevant in the future in high threat A2/AD environments and against the air forces of near-peer states, it must decide to go ‘all in’ on advanced drone and autonomy testing. Expert fighter pilots and engineers must work together to design software programs and run neural networks that put AI ‘pilots’ or agents through billions of slightly different combat scenarios so that when these autonomous drones actually go to war, they are able to handle or adapt to any situation that they may be confronted with. From a procurement perspective, UCAVs like Boeing’s MQ-28 Ghost Bat or Krato’s XQ-58 Valkyrie must be acquired in large quantities with multiyear purchasing agreements.

Designed to be a ‘Loyal Wingman’ aircraft, the Boeing MQ-28 Ghost Bat is a stealth, ‘fighter-like’ multirole UCAV, capable of flying and operating alongside crewed fighter jets. They serve as force multipliers and can significantly augment the combat power of crewed aircraft by performing autonomous missions independently through the use of AI and specializing in specific capabilities for a single mission. If acquired by DND, these drones will be deployed with Canada’s new F-35s and P-8A Poseidons as complementary forward sensors or EW and strike platforms, with some tasked with detecting targets, others serving as decoys or providing countermeasures, and others carrying out air-to-ground strikes along with air-to-air engagements.

Having said that, the MQ-28 Ghost Bat is by no means the only viable aircraft that can serve as a “loyal wingman.” With the ability to operate alongside crewed fighter jets or autonomously in its own swarm and carry missiles such as AIM-120 AMRAAMs both internally and externally, the XQ-58 Valkyrie is another low-cost tactical AI-enabled UCAV that the DND should seriously consider in terms of procurement as it would be a credible force multiplier for the RCAF’s future F-35s. The Kratos-developed aircraft can not only unilaterally accomplish missions such as scouting or serving as a decoy, but also execute offensive combat operations including electronic warfare support, SEAD, and offensive counter-air campaigns.

While UAVs are commonly known for their ISR and strike capabilities, they can also be utilized for air-to-air refuelling operations. A prime example of an uncrewed aerial refuelling drone is Boeing’s recently developed MQ-25 Stingray. If purchased by Ottawa, these autonomous platforms would unquestionably bring new functionality to the Air Force’s soon-to-be-acquired F-35s as they would almost double the effective combat range of these fighter jets, enabling them to travel 1,000 miles to their target area, conduct standoff missile strikes with sufficient dwell-time, and return to base. Such a capability is also especially key in vast areas like the Arctic where vital air interception missions are executed everyday to monitor and safeguard Canadian sovereign airspace.

The Stingray, however, in the event of a high-end war, brings more than its ability to autonomously deliver 15,000 pounds of fuel to 4 to 6 aircraft. It is also highly survivable as it features several stealth characteristics including a blended and rounded wing-body fuselage configuration, a horizontal shape absent any protruding vertical structures, a flush inlet to shield engine blades from radar, and a V-tail. Future plans for the integration of Long Range Anti-Ship Missile (LRASMs) and an electro-optical sensor turret on the autonomous drone are also underway, as part of an effort to expand its mission set.

LAND DOMAIN

While AI-enabled combat platforms can significantly enhance the power projection capacity of the CAF in any domain and serve as a force multiplier, their greatest asset is unquestionably their ability to help preserve the lives of Canadian soldiers. This capability is particularly vital for land forces as they will likely face the brunt of the fighting in any future major conflict. It is therefore imperative that the DND procure Autonomous Uncrewed Ground Vehicles (A-UGVs) such as Rheinmetall Canada’s Mission Master XTs which are designed to accompany and support troops during all kinds of missions, especially in challenging environments. With its robust mobility across extreme terrain and advanced amphibious capabilities, the Mission Master XT reliably transports payloads of up to 1000kg in even the most unforgiving weather conditions. It can also travel 750 km without refuelling and is robust enough to haul a platoon’s or company’s worth of sensors and weapon systems, greatly augmenting combat units’ range and firepower with heavier-calibre weapons and larger optics than they would otherwise be able to carry. Examples of fire support systems that could be fielded on this uncrewed platform include Fieldranger twin-M134 Miniguns, M2 heavy machine guns, Spike Non-Line-of-Sight (NLOS) missile launchers, and loitering munition launchers. Counter-UAS effectors along with tactical air surveillance modules, based on small AESA flat antennas radars, are also being considered.

Consequently, with such modularity, the Mission Master XT, if purchased, would enable the Canadian Army to undertake a variety of different missions in almost any situation without putting its troops at unnecessary risk. The operational spectrum of the A-UGV would encompass almost every aspect of warfare including supply and logistics, ISR, tactical overwatch, fire support, rescue, medical evacuation, CBRN detection, communication relays, and many other combat specific requirements. As if that were not enough, thanks to its compact footprint, this AI-enabled platform could also be transported inside a RCAF CH-147 Chinook or C-130 Hercules, and be delivered by parachute, to serve as a logistic or combat support vehicle for Canadian airborne troops.

However, while the acquisition of this uncrewed combat platform is unquestionably critical for Canadian infantry units, the procurement of autonomous light tanks to supplement the Army’s existing Leopard 2 fleet is equally important as it would bolster Canada’s armoured warfare capabilities without requiring significant amounts of additional funding and personnel. Such a purchase would be very prudent given the combat lessons learned so far from the war in Ukraine, particularly the growing vulnerability of main battle tanks (MBTs) against small cheap UAVs and loitering munitions, and the increased likelihood moving forward that future conventional armoured units will consist of both crewed and uncrewed platforms. This potential change in force structure is due to the fact that uncrewed AI-enabled tanks, unlike traditional MBTs, if destroyed, do not result in the deaths of troops. They are also cheaper and can be easily replaced.

As a result, given its current budget constraints, the Canadian Army must start looking at autonomous options on the defence market. Among the most advanced and viable platforms available is the Ripsaw M5 from Textron. This weapon system may be the ultimate solution and is an uncrewed ground combat vehicle intended to perform various missions including convoy protection, perimeter defence, surveillance, ground combat, and Explosive Ordnance Disposal (EOD) tasks. For ISR and navigation, it possesses and leverages a variety of systems including AI-enabled SUMIT 360 detection cameras which provide seamless Electro-Optical/Infra-Red (EO/IR) image composites for 360-degree situational awareness and threat detection, optionally tethered autonomous R80D SkyRaiderUAVs for persistent aerial surveillance, and Marsupial Uncrewed Ground Vehicle (MUGV) robots that can be deployed for standoff reconnaissance missions. It is also equipped with a TacFLIR 280-HDEP Medium Range HD Multi-Spectral Surveillance System that uses AI capabilities to rapidly detect and identify hostile targets at extended ranges, giving friendly forces a significant tactical advantage.

More importantly, in terms of combat capabilities, the Ripsaw 5 UGV can operate up to 1 km ahead of front-line troops and fire remotely using the Advanced Remote Armament System (ARAS), a gun that autonomously loads its own ammunition and can swap out various types of ammunition in just a few seconds. These capabilities would allow crewed vehicles to send the Ripsaw out in front of them and engage targets without exposing Canadian soldiers to enemy fire. Some weapon kits this uncrewed light tank can be armed with consist of a Mk44 Bushmaster II chain gun and a pair of CROWS-Javelin anti-tank guided missiles.

Lastly, but most importantly, an AI-enabled uncrewed derivative of the M142 HIMARS launcher vehicle, known as the Autonomous Multi-domain Launcher (AML), must be purchased by the CAF as a way to increase the Army’s capacity to conduct ground-based long-range missile strikes against ground and maritime targets without the need for large numbers of additional personnel. Such a procurement would be completely in line with the DND’s new defence policy which explicitly highlights the need to “acquire long-range missile capabilities to enable the CAF to deter threats to Canada” and “reach targets at greater ranges than our adversaries in combat.”

With regards to design, the AML is based on the M142, but is immediately visually distinguishable from the HIMARS vehicle due to its truncated cab section. More importantly, though, it is far more technologically advanced as it is designed to be able to follow a waypoint-based route autonomously, including as part of a convoy, from a hide location to a firing point. Once at a designated launch location, it can be remotely operated either by personnel immediately on-scene or at some distance using a “teleoperation” control method to execute long range fire missions. In terms of firepower, the AML is designed to use the same ammunition ‘pods’ as the HIMARS and is therefore capable of firing a variety of precision guided munitions including GPS-guided 227 mm rockets, ATACMSs and extended-range anti-ship PrSM ballistic missiles which have a maximum range of at least 1,000 km. Given this diverse arsenal of munitions and the HIMARS’ exemplary performance in the ongoing war in Ukraine, during which they have been heavily used to devastating effect against Russian ground forces, the combat effectiveness of a strike platform like the AML in a future conflict would be unmatched. For example, if war were to break out between the US and China and Ottawa were to get involved, the rapid deployment of these launchers along with crewed HIMARS to remote islands in the Indo-Pacific via C-130s or CC-177s would enable the CAF to effectively disrupt the PLA’s A2/AD capabilities and hold a variety of targets on land and at sea at risk. These combat assets, overall, if acquired, would give the Army significant strike capability against the Nation’s adversaries whether it be China in the Pacific or Russia in Europe or the Arctic.

OPERATIONAL CHALLENGES

Given all of these arguments, pieces of evidence, and proposals, it may seem that increased reliance on uncrewed AI-enabled platforms may indeed be the perfect solution to the CAF’s recruiting and retention issues. However, this new technology does not come without its challenges. It must be clearly understood that the use of autonomous assets does not mean that there are no personnel involved.These systems still require operating and repair crews to launch, recover, and maintain them. They may be uncrewed, but they have significant support teams that go with them. To make matters more difficult, these essential personnel must be highly skilled and proficient in electronics, computer software and hardware as well as cybersecurity. Such workers with those kinds of skills are hard to acquire and retain in the current Canadian labour market.

The military, nonetheless, still has a number of ways to overcome these challenges. For example, it can start divesting a large portion of maintenance routines historically done either by its personnel or maintenance facilities out to private industry partners. Such a measure would not only ensure the proper upkeep and modernization of the armed forces’ newly acquired autonomous combat platforms, but also serve as a potential new revenue stream for regional industries.

CONCERNS OVER THE DEPLOYMENT OF LAWS

Having said that, many defence analysts are still skeptical about militaries’ future willingness to actually field LAWS in large numbers due to these weapon platforms’ vulnerability to adversarial AI attacks, specifically data poisoning. Hostile forces and other kinds of adversaries have obvious incentives to manipulate one’s LAWS, and as soon as any single autonomous weapon is fielded, there is no way of knowing whether the weapon has, indeed, been already altered by an adversarial AI attack, or if it is being tampered with at that very moment. Data poisoning is a particularly dangerous scenario/example as it enables adversaries to be able to both reduce the overall accuracy/effectiveness of autonomous weapon systems by injecting corrupted data into their model’s training dataset and add “backdoors” to get enemy machine learning models to do exactly what they want undetected.

RESHAPING THE CAF

Overall, given the CAF’s growing personnel crisis and the escalating importance of faster ‘kill chains’ in modern day warfare, it is imperative that the military quickly evolve its force structure to consist of more autonomous uncrewed combat systems. Such an investment is critical to not only deter future potential adversaries, but also effectively offset the CAF’s ongoing personnel shortage problem which will likely persist for the foreseeable future. Furthermore, with an AI arms race already underway and the near certainty that hostile nations will continue to develop and test their autonomous weapons, Canada has no choice but to compete and build up its own arsenal or else be at an enormous military disadvantage.

Richard Nghiem is a Regular CDR Contributor