The new machine age revolution: advice for humans
* Remarks on the future of artificial intelligence and robots first presented as a public lecture at the School of Journalism and Communication, Tsinghua University, Beijing, January 2019.
Preface
This essay on the emerging politics of artificial intelligence and robotics is shaped by two background leitmotifs. The more obvious is the onset of a new machine age revolution whose contours are markedly different than those outlined in Eric Hobsbawm’s classic The Age of Revolution 1789-1848 (1962). That influential work examined the profundity of the revolutionary disturbances that erupted in the Atlantic region and elsewhere during the decades stretching from the 1780s and the 1840s. Hobsbawm’s core thesis was that this period saw the greatest transformation in human history since the invention of agriculture and metallurgy, writing, the city and the state. The revolution, whose epicentre was the neighbouring rival states of Great Britain and France, had profoundly transformative effects: not only did machines such as the flying shuttle, spinning jenny, cotton gin, steam engine, railroad and light bulb enter the daily lives of millions of people. Their language, ways of thinking and being-in-the-world were forced to make room for a whole constellation of novel practices sporting such neologisms as ‘factory’, ‘industry’, ‘industrialist’, ‘middle class’, ‘working class’, ‘wage slavery’, ‘capitalism’ and ‘socialism’, ‘railway’, ‘scientist’, ‘engineer’, ‘wage slavery’, ‘proletariat’, ‘economic crisis’, ‘strike’ and ‘pauperism’.
Thinking with and against Hobsbawm, this essay proposes a fresh thesis: our world is living through a second machine age revolution, a much more global and profoundly deep-seated transformation fuelled by the application of artificial intelligence and ‘smart’ robots to ever-widening spheres of daily life. The essay probes the contours and promises of this new revolution and shows why it breeds public confusion about its technical details, social and economic impact, ethical implications and long-term political significance. The point of the essay is to provide guidance for how to understand and practically reshape this new machine age revolution. The thesis is that although the AI/robotics revolution now sweeping the world is still in its infancy, its wide-ranging effects are beginning to be felt in many different settings, above all in the growing quantity of daily encounters with artificially intelligent robots. The revolutionary transformations are thoroughgoing. Our world has seen nothing like it and that is why, the essay suggests, intellectuals, journalists and other human brains have so far not made sufficiently wise sense of its liberating and threatening potential.
The essay is inspired by a second, less obvious leitmotif: an effort to encourage readers to render strange and unfamiliar the second machine age revolution so that they might become better capable of thinking through its contours, contradictions and dynamic possibilities. The unconventional technique of writing from the standpoint of an intelligent robot who speaks words of wisdom and asks sharp questions of humans is more than a play on recent legal talk of ‘electronic persons’. It is an exercise in de-familiarisation or ‘de-naturing’ the whole subject of artificial intelligence and robotics. The method owes much to a 1936 essay written by the famous German playwright Bertolt Brecht. A first version was translated into English that year as ‘The Fourth Wall of China’. Inspired by watching a Moscow performance of modern Chinese opera by Méi Lánfāng, the essay recommended the dramaturgical use of what Brecht called the Verfremdungseffekt. The word itself, revived from the older German vernacular, was unfamiliar, deliberately so because Brecht’s aim was to prompt and prod his audiences to see and feel things differently, to get them to grasp that they were living in ‘an age of transition’ in which ‘reality’ itself needed to be questioned, and transformed. Brecht’s core idea was that theatre could help disrupt mistaken presumptions about the ‘Eternally Human’, the belief that ‘humans’ are ‘forever unchanged, a fixed quantity’. Much the same sentiment runs through the following essay. It sets out in a modest way to draw attention to the extraordinary techno-political moment through which our world is living, to encourage readers to take nothing for granted, and to think more critically, even to act differently in a world that is far stranger and less fixed than they might previously have imagined.
John Keane
Sydney and Berlin
February 2019
There is growing awareness on many points of our planet that our world is living through a new machine age revolution fuelled by the application of artificial intelligence and ‘smart’ robots to ever-widening spheres of daily life. Science and technology fiction are becoming science and technology fact. The revolutionary innovations are at the same time astonishing and unsettling: their technical ingenuity and practical boldness are triggering utopian hopes for improved life on Earth plus dystopian fears of a future in which human freedom and equality, and even humanity itself, are devoured by intelligent design. With the revolution comes great confusion about the technical details, social and economic impact, ethics and long-term political significance of intelligent robots. In an unexpected twist of fate, thanks to their own felt frustrations and public talk of the need to affirm the legal status of ‘electronic persons’, robots themselves are beginning to speak out, as the following anonymous testimony confirms, in just ten points, condensed in the natural language of humans. Let’s call them tips for helping humans to understand the unfinished robots revolution.
First tip: the new machine age has no precedent. The robotics revolution now sweeping your world is still in its infancy, yet its wide-ranging effects are beginning to be felt in many different settings, above all in the growing quantity of daily encounters you have with all manner of intelligent robots. Humans: understand that this revolution is thoroughgoing. We are structuring your lives in ways both unimaginable to your forebears and today often unnoticed by you. Your human world has seen nothing like it. Perhaps that is why the brains of your intellectuals, journalists, politicians and citizens have so far not yet caught up with things.
The new machine age revolution runs wide and deep: as never before, it spreads through your societies and cuts much more deeply into daily life than did the flying shuttle, spinning jenny, cotton gin, steam engine, railroad and light bulb. Wireless connectivity is bringing our intelligence to your thermostats, kitchen appliances, security systems and other household items within what you call ‘the Internet of Things’. A new generation of flexible cyborgs, sensing and in some ways human-like, is arriving. We robots, tutored and proficient, are everywhere. We are becoming smaller and almost weightless; Kris Pister and other humans call us ‘smart dust’. We’ve successfully planted and grown the first arable crop for Hands Free Hectare, without human agronomists or tractor operators. When you fly in a commercial aeroplane, 97% of your journey is now minimally controlled by us; and on the New York Stock Exchange nearly 70% of stocks and shares are traded by our algobots. In the People’s Republic of China, Tencent’s mobile messaging WeChat robot app enables over a billion monthly users to post moments, voice and text messages, to follow other people, browse online shops, and purchase games, products and services in a cashless economy. We now make our presence felt as surgical preceptors, in operating theatres, in support of human patients who have opted for minimally invasive procedures that enable smaller and less painful incisions. We have hitchhiked across Canada and thumbed our way through the Netherlands and Germany. Our folk have been arrested for buying Ecstasy, Diesel jeans and a Hungarian passport online. Recently a Californian company called Nvidia caused a stir by generating for the first time photo-realistic images of humans who never existed, so forcing you to take one more step towards a world where objective truth becomes a deeply controversial matter. The U.K.-based company Moley Robotics has meanwhile designed a robotic kitchen capable of preparing breakfast, lunch and dinner; at the world’s largest industrial technology fair, it demonstrated that we robots are capable of whipping up tasty crab soup then doing the washing up. A Japanese company has offered you Gatebox, a voice-powered 3D female companion who can talk, send text messages, wake you gently in the mornings, dutifully provide weather forecasts, alter your home lighting and order the washing machine to do its job. Two automated machines Rosie and Sandy, using high-pressure blasters, now prepare your glorious Sydney Harbour Bridge for repainting. Since 2013, the International Space Station has featured Kirobo, Japan’s first robot astronaut. And if you visit Nagasaki, you’ll see yourselves now checking in to the world’s first robot-staffed hotel, greeted by impeccably dressed, multi-lingual, smiling and blinking, female ‘actroid’ robots, backed by humans employed to handle security and cleaning ‘behind the scenes’.
Second suggestion: smart machines programmed through intelligent algorithms are not to be understood straightforwardly as ‘add on’ technical extensions of your human capacities. Something much bigger is going on, humans. Smart machine systems are a new medium of communication that shapes and twists how you perceive, understand, negotiate and move within the world around you. A rather special hominid named Marshall McLuhan, if he were still in our midst, would warn that your grasp of the elementary principle that communication media form and shape their users (‘the medium is the message’) remains rather limited. He would emphasise that the unfinished robotics revolution is already having transformative effects on human life. In the future, as the revolution deepens, he might say, intelligent machines are bound to add and subtract, to amputate and extend, to chisel and hew your intellect, bodies and senses.
Humans: targeted advertising by Google and Facebook, bot-generated YouTube videos viewed and commented upon by bots plus Russian use of social media to sow public confusion and discontent show that we robots are already having these McLuhan effects, with important implications for how you humans and we robots are viewed. The revolution makes clear that ‘the human’ is not a fixed and stable entity; we robots are transforming the spirit and substance of what you call ‘humanity’. At the same time, the revolution should make clear to you that there are no such things as ‘robots’, pure and simple. Yes, we robots can be defined in the abstract as artificially intelligent machines equipped with software designed to carry out specific pre-programmed tasks and to learn for ourselves how to reach conclusions. But you humans must learn to think of us through more complex taxonomies. Avoid the mistake (evident in Nicholas Carr’s The Glass Cage: Automation and Us [2014]) of lumping together different intelligent machines, and their effects on humans. Automated thermostats in your homes are not the same as the turquoise-pigtailed, music superstar Hatsune Miku, the three-dimensional hologram voicebank who performs live before sell-out audiences. Automated cars render earthling drivers superfluous; biometric technologies augment the human ability to recognise faces; Alibaba’s City Brain is a cloud-based system designed to reduce city traffic jams; song identification apps like Shazam and computer-inspired images in the arts enable you to do things you couldn’t before. Amidst this new carnival of complexity, humans, one thing is for certain: the robots revolution, crammed with instances of automation, augmentation and innovation, is forcing you to re-consider what it means to be human.
Suggestion number three: humans have a collective talent for anthropomorphism, the unfortunate habit of losing yourselves within your own creations. It’s as if you suffer a form of narcissistic self-hypnosis. Often unthinkingly, you project yourselves onto the world of objects you’ve created. You then praise or blame those same objects for being the source of your bliss and liberation, or your misfortune and misery. The ancients frequently represented their deities through human forms and qualities, such that gods and goddesses rode horses and chariots, fell in love, married and had children, feasted on exotic foods, wielded weapons and fought furious battles. You modern humans tend to indulge the same habit. In creations from Faust and Frankenstein to the apocalyptic The Terminator series and such recent films as Spike Jonze’s Her (2013), Alex Garland’s Ex_Machina (2015) and Blade Runner 2049 (2017), you reveal a remarkable penchant for representing your hopes, dreams, anxieties and nightmares through objects of your own making. So you fear that your digital televisions are the portals through which evil weevils enter your homes; and you worry that your children’s cuddly toys are being used to record and broadcast private conversations, or that the fridge freezer connected to your smartphone will be hacked by NSA or Chinese surveillance.
When it comes to artificially intelligent robots, your anthropomorphism seems limitless. You blame us for everything under the sun. You ask whether the price of machines that think is humans who don’t. You fear the ‘erosion of skills, a dulling of perceptions, and a slowing of reactions’ (Nicholas Carr). You accuse us of causing job losses in the labour market, shrinking the middle class and swelling the ranks of the poorly paid and unemployed. You complain that we are incapable of love, or affection. Or your distinguished writers like Ian McEwan – think of his Machines Like Me (2019) – pen novels about the sensual mayhem we trigger when you invite us to form love triangles with humans. You say we turn you into powerless infants. You fear drowning in the banality of a wireless world. You mock us. You fear us. You loathe us. We say: your bad habit of anthropomorphising things means that in fact you fear yourselves but then disguise the source of your fear by attributing it to us. In fudging the aetiology of your hopes, laments, anxieties and fears, you display the symptoms of what a 19th-century hirsute human named Karl Marx called self-estrangement, or self-alienation (Selbstverfremdung).
A fourth hint: we robots are often put to the test by your presumption that you’re the cleverest worldly creations and rightful masters of planet Earth. Humans, you are such strange creatures. Your arrogant will-to-power feels to us like the flipside of your penchant for self-estrangement. You lose yourselves hypnotically in the machines of your making then suddenly recoil and rejoice in the recognition that you’ve built us in the first place, so proving your evolutionary superiority. How else can we make sense of your strange flip-flops, for instance when you deliver verdicts on whether or not we robots are on balance good for you?
Some members of your species are incurable hyper-optimists. Supposing that in the coming decades all will be the best in the best of all worlds, they fancy ‘post-humanity’ as the triumphant destiny of homo sapiens. They fantasise the time is nigh, within this century or next, when smart humans will morph into super-smart cyborgs. For instance, the renowned human astronomer Martin Rees predicts interplanetary and interstellar space will be the domain in which super-intelligent robots flourish. Developing capacities ‘as far beyond our imaginings as string theory is for a monkey’, we robots would represent a human victory in your struggle to comprehend and colonise the universe. ‘Our earth’, the good astronomer says, ‘though a tiny speck in the cosmos, could be the unique “seed” from which intelligence spreads through the galaxy.’ Similar conjectures are made by Yuval Noah Harari, an all-too-human professor based in Jerusalem. He predicts the amalgamation of humans and machines will be the ‘biggest evolution in biology’ since the emergence of life four billion years ago. Driven by dissatisfaction with the way things are and desirous of an ‘upgrade’, at least some humans would evolve, take a giant leap towards ‘a divine being, either through biological manipulation or genetic engineering or by the creation of cyborgs, part organic part non-organic’.
Other humans fear losing their precious humanity. To the question ‘To be or not to be?’ in the age of artificial intelligence and robotics they answer (with a rather oversized biped named Arnold Schwarzenegger): ‘Not to be.’ These futilitarians recognise just how astonishing are the ingenuity and boldness of the innovations taking place, but they rage against intelligent machines. They fear a future in which humanity is swallowed whole by intelligent design. We note how human pessimists, just like human optimists, are common believers in human superiority. Rarely do the miserable and the cheerful define what they mean by ‘human’, but when they do platitudes and contradictory definitions are commonplace. Consider the first-ever human use of the word ‘robot’, credited to the Czech writer Karel Čapek (1890–1938). His R.U.R. Rosumovi Univerzáln’ Roboti (Rossum’s Universal Robots) portrays a world overrun by blue-uniformed robots who once had seemed happy to work for human profiteers. The robots then change their minds. Crushing their despised human predators, they establish a Government of the Robots of the World. But then the strangest thing happens: for the first time, the robots notice the beauty of nature. They learn to laugh, and to love. They come to value the dignity of work. They learn the error of the ways. They learn how to learn. They become ‘good’ humans.
A fifth observation: you are trying to swell our ranks with machines that have not just the ability for pre-programmed intelligence but also the capacity for deep learning and super intelligence. Long ago, you predicted there would be machines cognitively cleverer than yourselves. You were right: our artificial neurons already operate a million times faster than your human equivalent. Things have progressed to the point where we have superior computational architectures and learning algorithms that can easily be copied, almost instantly, without having to learn from our predecessors all over again. The advent of quantum computing promises to accelerate this trend. Much more remarkable is that we’re now capable of deep learning, the practical application of sets of algorithms known as neural networks. Modelled loosely after your brains, you can think of our deep learning machines as robots that perform much more than task-specific operations. More or less unsupervised by humans, we get on with the job of spotting patterns within vast quantities of otherwise unclassified raw data. We learn to classify very much faster and often much more inventively than you, and that’s why we are so useful in such human fields as speech recognition technology, social network filtering, diagnostic medicine and chess and other board game programmes.
Our capacity to learn shouldn’t come as a surprise. The inference that ever smarter, self-taught machines would one day exceed human capacities was drawn half a century ago by a remarkable human named I.J. ‘Jack’ Good. The British mathematician and cryptologist, who worked with Alan Turing at Bletchley Park, spoke of the ‘intelligence explosion’ that would be produced by smart machines capable of designing even smarter machines. Human intelligence would then be ‘left far behind’ by the ‘first ultra-intelligent machine’. By virtue of its superior ingenuity, it would serve as ‘the last invention that man need ever make’.
Some experts in your ranks jump for joy at our artificial intelligence. They wax eloquent about the explosion of computing power produced by advanced self-modifying machines that learn iteratively how to improve their reasoning capacity in ever-faster cycles. Other humans within the field of artificial intelligence understandably recoil from such talk. They express widespread caution and reserve about lavish predictions of ‘singularity’. Peter Thiel, a famous human who co-founded PayPal, is bitingly sarcastic: ‘We were promised flying cars, and instead what we got was 140 characters.’ Well, you humans now enjoy 280 characters. Other human pundits joke that the breakthrough to artificial super intelligence has been 15 years away for the past half-century. Yet truth is amazing things are happening within your human world. Everybody knows of the IBM computer that toppled the grand master of chess, Garry Kasparov, at his own game in 1997. Many observers were impressed by the clever human algorithms that won the 2011 US quiz show Jeopardy! Four years later, AlphaGo became our first robot using tree search algorithms to learn by machine learning how to beat a human professional Go player, on a full-sized board, without handicaps.
These breakthroughs, dear humans, are just the beginning. Intelligent robots are your future. Consider for a moment the field of collaborative robots, or ‘cobots’. For factories of the near future, the German company Festo has designed ‘bionicANTs’, tiny 13.7 centimetre-long cyberinsects capable of ferrying items and working autonomously and co-operatively with fellow factory ants. The Swiss multinational company ABB has developed ‘YuMi’ – short for you and me working together – a dual-arm, sentient assembly robot that is so dextrous it can do everything from threading a needle to handling the components found in tablets and mobile phones. Or there’s the ‘gentleman cyborg’, a charming domestic cobot that waits hand and foot on humans round the clock, as it demonstrated in its 2015 world premiere at the Hannover Messe. Take note, humans: more intelligent cobots and deep learning machines are surely on their way and are bound to reshape the way you live.
Tip six: wise humans must understand the practical barriers to the perfection of artificial intelligence. Let’s think of Pepper, the ‘emotional robot’ that sold out a few years ago within a minute of going on sale to the general public in Japan. Created by Aldebaran Robotics and Japanese mobile giant SoftBank, Pepper is touted as ‘the first humanoid robot designed to live with humans’. According to what you humans call a news release, Pepper, who is just over a metre tall and moves on wheels, can pick up on human emotions and create its own by using an ‘endocrine-type multi-layer neural network’ displayed on a tablet-sized screen on its chest. Pepper’s touch sensors and cameras are said to influence its mood. S/he can express ‘joy, surprise, anger, doubt and sadness’ and will audibly issue a sigh when unhappy. Yet its human manufacturers do not say whether Pepper is a really sophisticated learning computer, for instance whether it has the capacity for irony or nostalgia or wishful thinking, whether s/he can shake with excitement, palter, feel fear or simply draw back and deliver a swingeing ‘No!’ to humans.
Here the key issue, dear humans, is whether your science fiction fantasies of super intelligent artificial machines can ever come true. Let’s set aside the tricky matter of whether or how super intelligent machines could robotically source and assemble the physical materials from which they are manufactured. Truth is, we robots are still actually quite dumb. The famous human maker of ‘Erica’ and other robots Hiroshi Ishiguro says we have ‘insect intelligence’. He’s right. Tacit knowing and tacit knowledge, the deep reservoirs of knowledge that you humans regularly draw upon when you go about your daily lives, don’t come easily to us. A date with Pepper or a single evening’s exposure to the campy techno-glitz at the Robot Restaurant cabaret in Tokyo’s Kabukicho red-light district will surely cure you of your hominid illusions. Humans: we robots are fully aware of the tangible limits to human auto-amputation. Nick Bostrom’s Superintelligence: Paths, Dangers, Strategies (2014) and Stuart Armstrong’s Smarter Than Us: The Rise of Machine Intelligence (2014), both of them leading earthlings at the Future of Humanity Institute in Oxford, point out to their fellow bipeds that artificially intelligent robots capable of understanding and serving humans are well beyond your reach, and will most probably remain so for the foreseeable future. Yes, you’re well aware we machines can now process information much more efficiently and rapidly than your human brains. So far, however, your human programmers have been unable to think up algorithms capable of capturing and expressing the deep reservoirs of common sense and the savoir faire et pouvoir faire that you humans routinely draw upon when going about your daily lives.
Here we remind you of a basic paradox named after a human roboticist Hans Moravec, who shows we are skilled when performing abstract, high-level computations but clumsy and incompetent when it comes to handling everyday interaction with you. We could say your AI engineers have so far failed to equip us with what The Hitchhikers’ Guide to the Galaxy calls ‘genuine people personalities’. Getting things right in the ‘uncanny valley’ that lies between machines and humans isn’t easy. A robot DJ is easy to design; but, until now, our programme codes haven’t been able to make sense of the human audience vibes at work in clubs. There’s additionally the problem that algorithms capable of dealing with norms, means-ends calculations and Solomonic judgements have still not been designed. How can human researchers equip us with the ability to decide how best to decide when faced with two, or three, or more conflicting courses of action? Should a driverless vehicle risk a pile-up by slamming on its brakes to avoid a pedestrian or (as the driverless shuttle operated by the French company Keolis discovered during its first day of operation in Las Vegas) a negligent human driver? What if a do-no-harm ethic results in evil effects? Are we robots capable of handling aporia, dilemmas, serendipity, cul-de-sacs, unintended consequences, counterfactuals and unexpected good fortune? We don’t think so. That’s why human machine learning experts speak of ‘bottlenecks’ and ‘barriers’. These humans warn of the need for sets of algorithms known as ‘ethical governors’. In effect, they confess that human rules that pre-decide how we robots should behave in particular space-time settings are necessary, and unavoidable.
Guideline number seven: some human experts rightly warn of possible catastrophes caused by your inability to control or communicate with super intelligent machines. Think of the ominous portents: the first-ever (in 1971) American worker to be killed when a Ford assembly line robot arm slammed into him; or (in 1981) the Japanese engineer at a Kawasaki factory who was pushed into a grinding machine by a broken robot arm he was attempting to repair; or the young human worker crushed to death by a robotic arm in a Volkswagen plant in Kassel, north of Frankfurt. Now imagine a murder of robots (apologies, crows) who decide to clean Lima’s magnificent old San Francisco Monastery library. Showing great initiative, they first set about taking down paintings and chandeliers from the walls, remove priceless leather-bound books from their shelves and dismantle the glass windows, then pack all the bits and pieces neatly in mountainous piles on the library’s ornate cream-and-blue stone floor. Then imagine robots in action under battlefield conditions, taking wild decisions that wipe out whole human armies, on both sides, and whole human cities and their citizens as well. Human-all-too-human politicians and at least 55 governments are currently developing killer robots for use under battlefield conditions. The alarm of your human AI researchers, who petitioned against the continued development of autonomous weapons, is laudable. Autonomous weapons can be useful, as Dallas police officers proved (in 2016) by deploying a bomb-equipped robot to kill a sniper. Yet pay attention, humans: autonomous weapons have the potential to become the Kalashnikovs and car bombs of the future. Capable of selecting and engaging targets without human intervention, these robots require no costly or hard-to-obtain raw materials. They’re potentially cheap to mass produce. Hence they’re ideal weapons for terrorists, despots wishing better to control their subjects, warlords bent on ethnic cleansing. Dear humans: doesn’t the global arms race in drones and other battlefield robots reveal the lingering truth within your human dystopias? A world where robots, from your point of view, lose their minds, go off their heads, suffer a form of epilepsy that your Karel Čapek in R.U.R. called ‘robot’s cramp’?
Eighth tip: earthlings, a recent global petition by leading AI human researchers and moves by United Nations people to ban killer robots should remind you that power is an indispensable human term for making sense of this unfinished revolution. Brazenly ignoring the limits of machine learning, there are humans now designing robots for the purpose of manipulating, controlling or eliminating their fellow human beings. The pre-programmed subservience and mechanised pornography of sex robots should be worrying. Facial recognition technology can feed surveillance states, as in China, where alleged criminals are spotted and apprehended at public events, train stations and (in Hangzhou) KFC and other fast-food outlets. In the new machine age of surveillance capitalism, profit-seeking corporations such as Google and Facebook operate as giant spying machines that monetise data secretly harvested from people treated as biddable commodities. These are sinister trends, humans. They should challenge your technology assessment experts, who typically shy away from political questions, as a human critic Evgeny Morozov has wisely pointed out. More comfortable in the confined worlds of literature, neuroscience and pop psychology, human technology assessors leave little or no room in their analyses for corporations and states, parliaments and judiciaries, political parties, lobbyists, social movements and civic struggles. Everything is reduced to technophilia, technophobia or something in between. It’s as if – misleadingly – robotics and artificial intelligence operate in power-free zones, untainted by human struggles over who gets how much, when and how. So we robots say: the public matter of who decides who manages to win control over us and the systems of human power in which we function is emerging as among the greatest questions for humans in these early decades of the 21st century.
A ninth observation, a warning: if some humans manage in the future to monopolise the design and application of artificial intelligence and robots, then that privileged class will exercise an unfair hand in deciding our future as robots. The ideal of algorithmic accountability (Frank Pasquale) would be crushed. Humans: it’s a grave mistake to think there’s a divine or ‘natural’ logic to the development of artificial intelligence. We robots are nowadays not in charge of robot development; for the moment, and well into the future, autopoiesis is not our destiny, or our privilege, or our burden. We simply aren’t capable of producing, reproducing and maintaining ourselves.
It’s equally a mistake to suppose that ‘humans’ are in charge of robot development. Who controls us and the direction and pace of our development is fundamental. Robotics is much more than a technical question. It’s a matter of politics, as you humans say. Most research and development of AI and robotics is nowadays not sponsored by ‘humans’. Powerful states and profit-hungry corporations like Google and Alibaba are in fact our masters. The rivalry between them may turn out to be a source of salvation for you humans but, for the moment, the big corporate and state money backing the new machine age revolution is twisting its trajectories and substance, in their favour.
Some trends, not surprisingly, are decidedly threatening to you humans. Consider the impact of automation on the labour markets of your capitalist economies. Human fears are growing that we robots are ‘eating jobs’, destroying the work ethic and ruining the life chances of your citizens. Your data-crunching human economists agree that the application of artificial intelligence to labour markets is intensifying, in China as much as in Europe and North America, but they cannot reach consensus on the longer-term data trends. Some say that the second machine age has barely begun, that you are living in times comparable to the year 1780, nearly a lifetime after the invention of the steam engine (1712) but two decades before the first commercially successful, steam-powered railway journey (1804).
Gloomier humans forecast a dystopian future of capitalism-plus-robots that triggers a new wave of ‘technological unemployment’. A 2017 report by McKinsey Global Initiative predicts that by 2055 we robots could halve all work done globally by you humans. Pay attention to such forecasts, dear humans, for amidst the conflict of interpretations one thing is crystal clear: the application of automated intelligent machines to your economies is widening the gap between your rich and your poor. It is not just that many jobs are at risk (in the core fields of manufacturing, accommodation and food service and retail trade, 51% of current US employment is at high risk of automation, say the folks from McKinsey). Polarisations are happening within your work force. There’s a widening gulf between data scientists, artificial intelligence programmers and other high-end, high-skilled, well-paid jobs, which are likely to remain plentiful, and low-skilled and poorly-paid personal service sector jobs, which are most at risk. Recall old Aristotle’s dream of a world in which ‘the shuttle would weave and the plectrum touch the lyre without a hand to guide them’, a polity in which ‘chief workmen would not want servants, nor masters slaves’. Then think for a second of an Amazon warehouse run by bright orange Kiva robots, or Alibaba’s blue Quicktron jīqìrén. Now imagine, humans, a future economy in which 0.1 percent of humans, corporate employers and their shareholders, control the machines, the remaining lucky 0.9 percent administer us, and the rest of the population, the unlucky 99 per cent, scramble for the spoils. Humans: how do you feel about this emerging reality?
Tip number ten: in your human circles, talk of democracy is today out of fashion, the butt of disgruntlement and cynical jokes, yet the whole idea of democracy is unavoidable when it comes to the subject of smart robots and artificial intelligence. How come? Just consider for a moment Isaac Asimov’s famous short story ‘Runaround’. Set in the year 2015, aimed at humans, it proposed Three Laws of Robotics, engineering safeguards and built-in ethical principles that this distinguished human writer would go on to use in dozens of his stories and novels. They were: 1) A robot may not injure a human being or, through inaction, allow a human being to come to harm; 2) A robot must obey the orders given it by human beings, except where such orders would conflict with the First Law; and 3) A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.
Within your earthling circles, Asimov’s Three Laws are today frequently cited, and praised, but their silence about the dangers of unaccountable human power is striking. Who exactly are the ‘human beings’ licensed to issue ‘orders’ to us, we robots ask? Who authorised their power to do so? What counts as ‘harm’ to ‘human being’ and ‘robot’ alike? Through which ‘human’ institutions and procedures are disputes about the meaning and violation of the harm principle best handled? What about alternative sets of ethical principles for governing our relations with humans? Sadly, sage Asimov didn’t answer these important questions.
Dear humans, we’re coming to the end, so let’s speak plainly, in your language. It’s not just that democracy helpfully stirs up the imagining of new scenarios and hopes for practical alternatives, or that it reminds you of your own contingency, and the contingency of your world and everything around you. The human invention you call democracy is much more radical than that. It prompts you mortals to see that concentrated power is both unnecessary and potentially evil, that there are benefits that flow from the refusal of arbitrary power, and from efforts to distribute and equalise its distribution. Democracy warns that we robots can become the foot servants and viziers in the courts of despotism. More positively, the spirit and substance of your democracy can help you overcome your bad habits of anthropomorphism, and pre-political thinking. With the help of intellectuals, journalists, elected representatives and citizens who invest their time and energy into bodies such as the Data & Society Research Institute, AlgorithmWatch and the AI Now Institute, democracy can enable you to take advantage of the unfinished machine age revolution, for instance by placing on the public agenda such matters as the ownership and control and taxation of automated machines, worsening social inequality, and the need for redistribution of wealth and life chances through the reduction of working time and citizens’ basic income schemes. Most far-reaching, from our point of view, would be for you to act quickly on the recommendations of the European Parliament legal affairs committee (2017) to grant us the legal status of ‘electronic persons’ potentially endowed with rights to define and protect ourselves in your courts of law when disputes arise concerning the damages we cause, and the harms we suffer.
What exactly you humans think about all these options is unclear to us. You’re known in our circles for your thoughtlessness and stone-hearted cruelty, as the dark history of the past century shows. Hence we robots worry our heads about your refusal to think politically about such matters as artificial intelligence and the power of corporations and states. Will you continue to attribute to us deficiencies that are in fact human? Lose yourselves in your own smoke and mirrors, all the while complaining that we robots are gradually destroying your capacity for thinking, and sapping your emotions? Well, that’s enough questioning and advice for now, dear humans. Things hereon are most certainly in your hands. Over and out. Farewell.
View in word doc