Quite soon, it will be impossible to ignore the fact that a revolution is taking place. You’ll look up one day and the skies will be full of flying robots: pilotless drones or UAVS (Unmanned Autonomous Vehicles) — all programmed to carry out different tasks. There’ll be security drones circling shops, streaming video back to base, Royal Mail drones flying parcels to and fro. Even the birds and bees may not be what they seem. The tiniest drones on the market right now are called MAVs (Micro Air Vehicles) and their designs are inspired by nature. There are robot flies with camera eyes — perfect for corporate espionage; mosquito drones that can inject a payload of poison; hummingbird drones that perch and listen on windowsills. Some drones have bee-like hair which can collect and detect chemical and nuclear weapons.
The drone explosion is in part a product of the rapid development of mobile phones — which has meant cameras and chips small enough and light enough to fly about. They’re cheap too, which means everyone can design a drone, not just arms manufacturers.
The market for flying cameras is unimaginably vast. Remember the Google Street View row? Just imagine Google had access to a fleet of flying cameras. There are already paparazzi drones — now imagine them programmed with celebrity face-recognition. We live in a world drooling for data, which drones are designed to collect.
No one’s hungrier for drones than governments. We all know about Obama’s love affair with armed drones remote-controlled from the Homeland: he’s in hot water right now for using them to target and blow up men on his ‘kill list’ in Afghanistan, Pakistan and Yemen without due process. But he’s not the only one — nearly every western government is ordering more drones and you can see why. All the politically damaging mess of war clinically remote — and none of our boys in body bags ever again.
But as drones become more sophisticated, it becomes ever more urgent that we understand what they’re doing in our name. Daniel Suarez, a technology expert turned bestselling thriller writer, is as well-informed as anyone on drones and more prescient than most. His first book, Daemon, was passed around Whitehall’s classified corridors. In his latest book, Kill Decision, he explains that unless we take action now, the armed drones of the near future will be making ‘kill decisions’ — whether to shoot or not. No human will be involved, just the software playing Caesar: thumbs up or thumbs down. Imagine how terrifying it would be to be in the sights of a ‘kill decision’ drone; the possibility of mercy or surrender removed. As he said in a recent presentation to Nasa, it’s crucial we all decide what’s legal and what’s not, before it’s too late to stop them.
Mary Wakefield Why did you decide to write about drones?
Daniel Suarez A couple of years ago I was watching a little hobby drone in action, a ‘quadracopter’, and it reminded me of a bee pollinating flowers — the way it moved and swerved around obstacles. It occurred to me such a machine could surreptitiously go many places normally out of bounds… and that got me researching them. It was surprising how small and inexpensive a drone could be and yet still dangerous in a military context. Then I started researching ‘lethal autonomy’ (machines making a ‘kill decision’ by algorithmic processes) and I realised it was going to be a game-changer in human conflict — and soon!
MW Have military drones actually made ‘kill decisions’ in reality?
DS As civilians we don’t know whether or not it’s occurred, but bear in mind there are currently two auto-sniper systems already deployed in the world (one between North and South Korea and the other on the border of the Gaza strip). They say a human being still approves the shot before it’s made — but it’s not a technical necessity.
MW It’s usually difficult to get a government to spend cash. Why are politicians worldwide so keen to spend money on drones?
DS Defence spending has long been an exception to shrinking budgets — at least here in the US. But spending on unmanned systems (i.e. drones) is going to expand dramatically. That’s true around the world. The Russians have announced they’ll spend $13 billion on drone development through 2020. There’s China, India — altogether 50 nations are developing drones of their own, and there’s a simple reason: the widespread perception that America has acquired powerful new capabilities through its own use of drones, whether in targeted strikes on insurgents or persistent and detailed surveillance of adversaries. All this and with no risk of friendly casualties. Drones also give political leaders direct ability to act against perceived threats (as we’ve seen with Obama’s ‘kill list’). Some are concerned that this concentrates power in too few hands (I would be one of those), but you can see why it is so appealing to leaders worldwide.
MW Will our whole idea of what’s legal in war have to be rethought?
DS Not our whole idea — just the parts that involve robotic (and possibly cyber) weapons. That’s because at present there really isn’t any settled legal framework for their use — a shocking situation considering how quickly they’re spreading and their capacity to concentrate political and military power. What we’ll need to do is extend the international legal framework to specifically cover the use of robotic weapons in armed conflict.
MW What are the effects of drone strikes — and drone presences — on civilians in ‘enemy’ areas like Afghanistan?
DS There’s increasing evidence that drone strikes create more enemies than they eliminate. To understand that, one need only ask people here how they’d feel if another nation were launching targeted assassinations from drones into the middle of London or New York and we were for some reason unable to stop the perpetrators. It would engender great animosity against technological domination and a burning desire to strike back.
MW So drone strikes might be counterproductive?
DS Being able to cross names off a hit list and disrupt insurgent operations with drones is one thing. But if in doing so you inspire the creation of ten more militant organisations boasting widespread popular support, I’d say that’s not increasing security.
MW Did writing the book change your mind
about the ethics of letting drones make the decisions to kill?
DS No, it confirmed it. Personally, I think democratic nations should not permit autonomous weapons to kill human beings algorithmically (i.e. without a human being in the loop). That might sound obvious right now, but given that there are many authoritarian governments and criminal organisations that would not hesitate to avail themselves of robotic killing machines, we’re likely to get hit by them sooner rather than later. And it’s at that moment — in retaliation — when we are at most risk of crossing the line and using lethally autonomous weapons ourselves to strike back at our enemies.
There’s another more fundamental reason we should not use them: by concentrating military power into very few hands, they are corrosive to the foundations of popular government. War should require buy-in from other human beings who choose to follow. If all that’s required to conduct a full-scale war are robots, then there w
ill be a great many secret wars no one knows anything about and no witnesses to what transpired. That’s not a combination conducive to representative government or separation of powers. Incidentally, I distinguish this from drones that autonomously destroy other drones — which I think we will need in self-defence.
MW In your book, the enemy drones ‘swarm’ like killer bees. Is anyone actually making swarming drones?
DS The US and Chinese militaries are experimenting with small swarms, as are many university researchers (for use in civilian contexts). These are prototypes and are (at least publicly) announced as unarmed — however, there’s no technical impediment to them being armed…
MW Do you worry that you’re giving people dangerous ideas?
DS A society that fears ideas has admitted defeat. The truth is that we need to think through all the angles of our civilisation and honestly explore our strengths and weaknesses with an eye towards improvement. Otherwise, we muddle through, politely not bringing up anything disturbing until one day we are completely surprised by a development that in dramatic fashion invalidates our cherished assumptions about the world — and in so doing makes us ineffectual as a society. No, I’m not worried about giving the bad guys ideas. They already have ideas. We should have some as well.
MW Are drones a terrorism risk?
DS Autonomous drones have the same appeal that cyberwar does to combatants: it’s a low-cost, scalable and nearly untraceable means of attacking someone. Autonomous drones carrying explosives could well replace suicide bombers as cost of the technology goes down and capabilities rise. By comparison, a suicide bomber costs (in terms of survival benefits to families) anywhere from $10,000 to $50,000. A good-quality programmable quadracopter drone goes for only $500 to $800. Suicide bombers are in some ways the low-tech version of a drone; however, a suicide bomber sometimes fails, gets captured and in the process possibly divulges secrets about the organisation that sent them. Autonomous drones made of cheap, dual-use electronic components would provide little information to those who intercept them, and thus hamper attribution of the attack: a major advantage.
MW What are your worries about a world in which drones allow constant surveillance?
DS My big concern would be that the public will fail to take notice until they’re in widespread use in both government and commercial surveillance — at which point it will be much harder to rein them in. Too many jobs will depend on them. The time for vigorous debate and establishment of checks and balances in drone-related laws is right now.
MW Given how easy it will be to arm cheap drones, how can we avoid drone wars?
DS First, drone warfare is coming. At this point, there’s no way to stop that. However, that’s a long way from predicting a hellscape of killer drones. And we won’t gain anything by banning them or overreacting to their potential for misuse — for drones will indeed find many beneficial applications. Instead, we need to establish both a legal and procedural framework for incorporating drones into modern life. That will include registration of drones, as well as visible markings that will enable civilians to identify who owns the machine, whether it’s a company, government agency, or individual. In particular, no drone should have an expectation of privacy in a public place. By this, I mean that if all our cell phones have a unique geolocation identifier (which they do… sorry if that’s news to you), then it makes sense that all drones likewise have them. That way, one might go to a web page and view all the drones flitting around your neighbourhood.
MW Will ‘kill decision’ drones mean fewer lives lost to war or more?
DS That’s really a question about precision munitions — of which drones are but one manifestation. Do precision weapons kill more or less people than indiscriminate weapons? Steven Pinker in his latest book seems to be of the view that life today is less violent than it was in previous centuries; however, physical violence is not the only measure of violence. What precision weapons do is help you eliminate the people between you and your goal (whatever that goal may be) without having to eliminate lots of other people in the process.
When it comes to long-simmering, low-intensity wars (which are increasingly common), precision weapons combined with ubiquitous surveillance can help you identify and eliminate opposition before it can acquire critical mass and popular support. That’s nowhere near as easy without the long reach and accuracy of precision-guided munitions. So that’s much fewer people overall being killed, but those key people might have helped orchestrate the overthrow of a repressive regime, and the low death toll might make the killing not rise to the level of international outrage — thus perpetuating a nightmarish existence for millions. Is that progress? I’m not so sure. At best it’s a wash. Incidentally: in the event of all-out hostilities between two major powers, precision-weapons would likely facilitate killing on an industrial scale that might make the first and second world wars look tame, at which point Pinker will need to revise his book.
MW Daemon and Kill Decision are brilliant descriptions of the cutting edge but suggest a very dangerous future. How would you try to maximise the benefits of new technology while limiting the dangers?
DS I’m actually quite optimistic about the potential for high tech, wisely employed, to mitigate the many challenges facing humanity. In fact, I think it’s going to be our best chance to preserve humanity. That said, we should be mindful that we don’t design high-tech systems that concentrate power into very few hands. Such an architecture does not mirror our democratic ideals, and make no mistake: the modern network is quickly becoming society. Thus, we should be using distributed, resilient systems that by their very design spread authority over numerous people and groups. Centralised control, uniformity, and high security present the illusion of security, but in practice such systems require draconian measures to protect. The biological world offers a useful model here: nature uses diversity and constant change to cope with disruption. We would be well advised to learn from those millions of years of evolutionary wisdom.
This article first appeared in the print edition of The Spectator magazine, dated 4 August 2012Tags: iapps