Robots, War and the Future of Airpower

In the near future air operations will happen at altitudes as low as the heel of a soldier’s boot as aircraft capable of offensive and defensive operations reduce in size and increase in range.

That this is likely is demonstrated by the slaughterbots video that went viral in November 2017.

The slaughterbots are autonomous offensive drones that can fit in the palm of a human hand. Fitted with explosive warheads designed to impale human skulls, these very low altitude drones were depicted in the video as having the ability to autonomously select and engage targets based on age, sex, fitness, uniform and/or ethnicity.

The video ended with one such robot killing an unarmed student lying under a desk in a lecture theatre (a clear-cut war crime). Legitimate militaries have no interest in perpetrating genocide, but the prospect of autonomous drones attacking human combatants with explosives and projectiles is very real.

Very Low Altitude Drones

Future weaponised micro-aircraft will be capable of operations at very low altitudes – just millimetres above the ground. Indeed, some aircraft may become hoppers having the ability to fly through the air and drive on the ground.

This descent to ankle altitude of air power will fundamentally change the nature of army operations. Increasingly, they will come to resemble air force operations. Effective air power is now small enough to fit in a backpack or armoured personnel carrier. This low-altitude air power is mostly used for reconnaissance enabling an infantry unit to have an eye in the sky; however, these systems are starting to become weaponised.

In future high-intensity war, infantry units will need to be familiar with air force logistics and tactics. Future ground troops may need to be airmen as well as soldiers in that they will need to understand low altitude air war in addition to land war. Alternatively, combat airmen may need to be embedded in infantry units.

One might argue this is the continuation of a process that started when the cavalry switched to helicopters from horses. Regardless of the institutional arrangements, ground units will need defensive drones, portable anti-aircraft weapons they can carry themselves, or both.

Tiers of Air Superiority

In the near future, it may be there are tiers of air superiority. At low altitudes, the air will be dominated by large numbers of relatively small and cheap aircraft flying at low speed and having relatively low range. At higher altitudes, existing air force concepts of air superiority will continue to apply.

There will still relatively small numbers of expensive supersonic fighters clearing the skies of enemy aircraft but F-35s burning through the sky at high speed will be of limited use against swarms of slaughterbots flying less than 10 feet above the ground at relatively low-speed attacking infantry and other ground-based targets.

Indeed, already we have seen how such cheap low altitude drones can be used to attack expensive hi-tech aircraft sitting on the ground. In Syria, in January 2018, a low-tech attack by Syrian rebels was made on hi-tech Russia on its airbase at Hemimim and its naval station at Tartus. Militaries will find it irresistible to develop weapons costing a few hundred dollars that could destroy supersonic aircraft worth tens of millions on the ground.

Naturally, countermeasures to guard against such threats have already been developed. The Russians were able to see off the attack.

As an aside, this gives the lie to the claim made in the slaughterbots video that ‘humans will have no defence’ against such attacks. In the history of war, the measure has always led to the development of counter-measures. The spear led to the shield. The submarine led to the depth charge. Barbed wire and machine guns led to tanks.

The slaughterbot will lead to counter-measures just like every other technological innovation in military history.

Traditional High-Altitude Air War Will Still Be Critical

Even given the new low-altitude threats, F-35s and other fighter aircraft will still be able to target the transport planes and vans delivering the low altitude ‘slaughterbots’ to a range close enough for them to attack in a conventional war. Destruction of enemy logistics will still be a key to victory as it was in the Battle of Midway and many other historic engagements. However, given suitable terrain, it is conceivable that one belligerent in a future high-intensity war might have air superiority at altitude while the other has air superiority close to the ground. 

Increasing Autonomy

Even in an asymmetric war, a plucky low-tech belligerent might find a way through the emerging countermeasures and achieve low-altitude air superiority with devastating effect to a high-tech foe. We can expect those who perpetrated the failed attack on the Russians in Syria to go back to the drawing board and try again, this time targeting the low altitude air defence systems first (perhaps with a suicidal or stealthy ground attack) before unleashing the drones on the high-value targets in the hangers.

Increasingly these low altitude aircraft will be autonomous in their combat functions. This is because such craft is far too small for an onboard human pilot. Also, in the near future, a fight between a human-telepiloted Uninhabited Air Vehicles (UAV) and an autonomous UAV will be as fair a fight as Kasparov vs Deep Blue or Sedol vs AlphaGo.

There will come a time where the AI has advanced to a point where humans cannot defeat it. A further reason is that existing drone counter-measures use techniques such as jamming GPS and telepiloting frequencies. To counter the counter-measures, drone-makers could resort to dead-reckoning or visual navigation to avoid the GPS vulnerability. To avoid the telepiloting vulnerability, they could disconnect the network card and develop onboard autonomy. 

Moral Arguments regarding Lethal Autonomy

There is, of course, a moral argument here in addition to the technical ones regarding lethal offensive autonomy. It is foreseeable that there will be a ban on autonomous weapons but terrorist and criminal groups (narcoterrorists) will readily adopt such low altitude air power because of its low cost and easy availability. Indeed, they already have.

ISIS weaponised telepiloted hobby drones in 2016. Drones are now being used to smuggle drugs across borders. It is only a matter of time before cartels use them to assassinate police, judges, and ministers. This will force democratic nations to adopt counter-measures.

To be effective, such counter-measures, like existing close-in weapons systems such as Phalanx, will need a high degree of targeting autonomy. Historically, military necessity has trumped moral objections to many new weapons because belligerents will do whatever it takes to win.

Jefferson Davis objected to ‘torpedoes’ (mines) as ‘cowardly’ weapons, but when the Confederacy got desperate, he acquiesced to their use. Submarines and bombers were similarly objected to on moral grounds yet because they were militarily effective and could be used in compliance with principles of discrimination and proportionality, they survived.

Military norms changed to accommodate them. The Campaign to Stop Killer Robots may get a ban on autonomous weapons, but these efforts may not bear fruit.

Already, the Russian military has prototyped an autonomous tank, the Neretha, which was reported to have outperformed human tank crews in recent trials. A US team of researchers have already built an air combat AI that defeated a retired Air Force colonel in a tactical simulation.

Autonomous Swarms

It is likely autonomous low altitude aircraft will function in swarms and be teamed with humans who will increasingly focus on defining military goals at a high level rather than figuring out how to achieve them in detail.

A human officer might say to a swarm of robots, ‘clear that house’ and the robots will autonomously work together to search for threats and clear it. Swarm robotics is a rapidly progressing area of research. If the swarm of robots needs to shoot enemy combatants therein (be they human or robotic) they will do so within the scope the human order “clear that house” gives them.

Similarly, having disposed of the low altitude air defence, a commander could say to his swarm ‘search for and destroy all Sukhois on this airfield’, and the robots would do this. If they needed to shoot airmen trying to defend their craft, they would do so within the scope of their human order and the normative constraints of targeting law.

At present, such human-robot teaming is not practical, but it is a defined research goal. Most drones remain telepiloted; however, voice-controlled robots that are intelligent enough to respond to simple tactical instructions are not far away, and autonomous robots intelligent enough to comply with targeting law are not so far away either.

Such airborne robots would integrate with ground-based robots. Ground-based robots will be primarily logistical with strike ability mostly airborne except for some very particular tasks, such as sneaking up like snakes on enemy low altitude air defence posts. Command and control will be distributed and redundant. Cybersecurity will be critical.

Much research and development will be needed to ensure these drones operate in accordance with targeting law. However, such research is being funded. In the absence of a policy prohibition, the advanced powers will succeed in developing normatively compliant autonomous weaponry. In a climate where major powers do not trust each other, each will keep their guard up.

Increasingly, given the advances in AI and robotics, the front line of combat will be a combination of Uninhabited Ground Vehicles and UAVs on land and Uninhabited Sea Vehicles, Uninhabited Underwater Vehicles and UAVs at sea.

Future War May Be Predominantly Robotic in the Front Line

Future war may well evolve into robot vs robot at the front line. Human resistance against robots may become futile. Should such conditions evolve in war, it may be conventions evolve to take human surrenders in such hopeless circumstances.

Indeed, future war might become as ‘civilized’ as the wars of the Italian condottieri in the Renaissance. Machiavelli wrote of the Battle of Zagonara in 1424 that it was a ‘great defeat, famous throughout all Italy’ and yet ‘no death occurred except those of Lodovico degli Obizi and two of his men, who having fallen from their horses were drowned in the mud.’[1]

Historians doubt the body count in Machiavelli’s report as he had it in for the lack of warfighting prowess of the condottieri, but even so, just as the wars of the condottieri were more about manoeuvre and posture than actual hard fighting, future war might become a matter of destroying material rather than people.

Some would argue war has been as much as about destroying material rather than people since the Second World War as evidenced by the raids on Schweinfurt and Bologna and the U-Boat campaigns in the North Atlantic.

Once the opposing robots are destroyed, the humans may surrender as the ability of humans to defeat robots in combat might no longer exist. Resistance without robots may turn out to be as futile as trying to beat Deep Blue at chess without a computer.

In a future high-intensity war, it might be written:

“[u]sing a new technological invention, the Red robots wiped out the Blue robots in the first days of the war, compelling the Blue humans to surrender. Some Blue humans tried to fight on, but the Red robots disarmed them, laughing at their slowness. The videos taken by the Red robots went viral.”

Sean Welsh is the author of Ethics and Security Automata: Policy and Technical Challenges of the Robotic Use of Force and a postgraduate student in Philosophy at the University of Canterbury.

Prior to embarking on his PhD, he wrote software for British Telecom, Telstra Australia, Fitch Ratings, James Cook University and Lumata. Sean tweets as @sean_welsh77

[1] Niccolo Macchiavelli, History of Florence, Book IV, Chapter I.

This article was republished with the permission of The Williams Foundation and was first published in their column The Central Blue.