Imagine several terrorist drones carrying primed hand grenades buzzing towards a civilian or military target, leaving its defenders with just moments to deploy their state-of-the-art countermeasure. With a furious beating of wings, the precision weapon designed to send the deadly machine crashing to the ground is launched – a squadron of kamikaze pigeons.
Far-fetched as it may sound, the idea is one of those being explored by American researchers as the defence and security sector scrambles to respond to the extraordinary pace of development in drones – or Unmanned Aerial Vehicles (UAVs) – which is seeing robotic and increasingly capable autonomous aircraft move rapidly from the purview of state militaries into the technological and commercial mainstream.
The pigeon research, being carried out by an unnamed private US company according to an industry expert, aims to harness the birds’ natural flocking instincts and acute vision to draw them towards the particular sound and motion of a “quadcopter” drone. The birds will be trained to fly at a machine or machines en masse, neutralising an attack on a potential target from an airport to an open-air concert at the cost of their own lives.
Although currently experimental, the idea is not unfounded in reality. During the Second World War, US scientists perfected a technique to train pigeons to effectively pilot a glider-type bomb by conditioning the birds to recognise an image of a target. More recently, birds of prey have been trained to intercept drones entering their territory, albeit with mixed results.
The “suicide pigeons” sit alongside an ever-increasing and bewildering array of Unmanned Aircraft Systems countermeasures (C-UAS in the industry jargon) being offered to governments and infrastructure operators, ranging from nets fired from missile launchers to ultra-sophisticated radars, lasers and jamming devices designed to interfere or destroy drone guidance systems. The value of the C-UAS market is predicted to grow by nearly a third every year for the next six years, reaching £3.4bn by 2027.
But the very fact that something as idiosyncratic as self-sacrificing birds is being seriously investigated as a means of thwarting drones is equally strong evidence for a more sobering proposition – that offensive drone technology is evolving at a rate which poses a considerable threat to even the most sophisticated militaries. It also poses a particular threat to civilian populations if the technology is mastered by terrorist groups intent on targeting western countries.
Security sources told iweekend that the evolution of so-called drone “swarms” – the use of multiple drones working in an interconnected fashion using artificial intelligence (AI) – for military purposes will be one of the single-most potent challenges created by the next generation of weaponry.
As one UK official put it: “Everyone saw the drone display at the opening ceremony for the [Tokyo] Olympics. Now imagine those 2,000 or so drones not making pretty pictures but each armed with a smart munition and programmed to seek multiple, interlinked and changing targets – such as an infantry battalion, or even just a crowd. And then imagine that you don’t even need multiple operators of those drones. They are autonomous and you just need to press a button to launch the equivalent of an entire air force. That is the sort of capability that is fast coming around the corner.”
It is a point that was underlined this week by General Sir Patrick Sanders, the officer in charge of overseeing the strategic direction of Britain’s defence forces. He he told an audience at the world’s largest arms fair in London’s Docklands that China is seeking to establish a global lead in technologies including “autonomous swarms”. At the same time, the Ministry of Defence announced trials by the Army and the Royal Navy of Britain’s first C-UAS laser weapon designed to shoot drones from the sky.
Military and some academic experts caution that the prospect of such a fearsome weapon as a militarised drone swarm being developed or acquired by a “non-state actor” is currently remote.
But others, including one of the world’s leading AI experts, argue strongly that time is running out – if it has not already done so – to prevent the proliferation of autonomous or self-controlling swarms which not only fly themselves but take the decision on whether or not to open fire.
Stuart Russell, professor of computer science at University of California, Berkeley, said the large-scale manufacturing of lethal autonomous weapons was a “nightmare scenario”.
He told iweekend: “There are tens of millions of AK-47s in non-state hands, and we would expect tens of millions of small lethal drones to end up in non-state hands as the scale of manufacturing increases and prices go down.
“The big difference is that a million AK-47s require a million human soldiers, whereas a swarm of a million lethal drones requires a single human operator. So, it is not unreasonable to expect that at some point we will see mass attacks, maybe in Tel Aviv, or Paris, or Washington DC.”
The British-born academic is one of several experts to point to events in the hostile deserts of western Libya last March, when a fleet of Turkish-built Kargu-2 drones were launched in a battle between factions vying for control of the oil-rich state once ruled by Muammar Gaddafi. According to a United Nations expert panel, the machines may have been programmed to hunt down and attack a set of retreating opponents without human intervention.
The 548-page study on the conflict in Libya, presented to the UN Security Council in March, concluded: “The lethal autonomous weapons systems were programmed to attack targets without requiring data connectivity between the operator and the munition: in effect, a true fire, forget and find capability.”
STM, the drone’s manufacturer, has insisted that the machine – which has been sold to several unnamed militaries – does not have a fully-autonomous capacity.
Experts have expressed concerns that manufacturers in “rogue states” are unlikely to share similar scruples.
In the meantime, a more immediate threat comes from weaponisation of commercially-available drones. Rebel groups in Syria, as well as in Iraq and Afghanistan, have been shown to have routinely modified quadcopter drones made by manufacturers, including China’s dominant DJI, to allow them to carry and drop munitions with devastating effect against their opponents. Counter-terror experts have repeatedly warned of the risk of deploying similar tactics away from war zones in European and American cities.
In 2018, 13 drones were simultaneously launched in a rebel attack on a Russian air and naval facility in Syria. While the assault did not use “swarm” technology, experts argue it was a sobering example of the sorts of tactics already available to those with malign intent. Indeed, the ubiquity of drone use by terrorists and criminals was underlined by the Christchurch mosque shootings in New Zealand in 2019 after it emerged that the killer had deployed a small drone to reconnoiter his target, recording images of exits and entrances from the air.
Zak Kallenborn, a leading expert on terrorism and emerging technologies, told iweekend: “Massed drones allow terrorists to strike many targets at once or overwhelm the defences of a target. If a terrorist attempted an assassination with 20 drones, only one needs to be successful. But defenders need to be successful in defeating every one. Off-the-shelf drones are also increasingly autonomous, so the terrorists launch the drone and escape before the first bomb explodes.”
The result is a two-way arms race with countries racing to develop or acquire the most advanced drone systems and at the same time the means – whether it be squadrons of kamikaze pigeons or laser guns – to defend themselves against an entirely new class of weaponry. As the UK security official put it: “The question is whether you can defend yourself in all circumstances and at all times. History shows that is a difficult thing to achieve.”
From attack dolphins to incendiary monkeys – animals have been used as weapons throughout history
The use of animals as sacrificial pawns in human combat is, grimly, not a new idea. Throughout the Cold War, the Soviet bloc trained dolphins for use against human and naval targets.
The marine mammals, based at a top secret marine unit in Sevastopol, were taught to intercept enemy divers with harpoons and carry out “suicide” missions against vessels with explosives strapped to their backs.
The US also trained dolphins for military use, though Washington insists its programme did not use the animals to harm humans but instead to guard naval facilities and use equipment to snare intruders.
Several other countries, including North Korea and more recently Iran, are said to have looked to use marine mammals in a conflict situation, with Tehran reportedly purchasing trained animals from Russia.
The use of animals as weapons is thought to go back to 10th-century China, where manuscripts record the use of monkeys – encased in straw and set alight – as incendiary devices to be unleashed on enemy camps.