On December 5, 2018 I was honoured with an invitation to open the “Fantastic Futures AI Conference” organized by the National Library of Norway, in Oslo. I put together my talk like a triptych of words; “Work”, “Love” and “Polis”. For each of the words I used a story to explore what the words mean – or might mean – in a world of intelligent machines. The text below is the transcript of my talk:
In my first story a wise and omnipotent Creator decides to fashion two creatures in His Own Image, a male and a female. I guess you know their names…
He places them inside a beautiful garden where everything is plentiful, so the two happy creatures have nothing to worry about. They can spend their days doing little else but enjoy themselves, bask in the eternal sunshine. It’s a world of material abundance and endless leisure. Sounds like a Paradise – right? Anyway, something happens, the female gets somewhat curious about something that was out of bounds – the Creator gets really pissed off about that, and that’s the end of the short holiday we humans enjoyed in the Garden of Delights.
As our great-great grandparents get kicked out a heavenly, thundering voice delivers, by way of punishment, some fateful words: In the sweat of thy face thou eat bread! In other words: If you want to eat – go get a job!
Some say that this biblical story encodes a deep trauma in human evolution. The pain and anguish of transforming from a free hunting and gathering species into a hard-toiling, face-sweating, farming species, subject to laws made by kings. That the story marks beginning of the agricultural revolution. The story also marks the genesis of economics: the beginning of scarcity. It’s all about work from now on, and money, and power. And war too – for isn’t war the result of scarcity?
The story of Adam and Eve led many Christian faithful to uphold work as a way of redemption. But even the unbelievers among us share a deep-rooted belief in the intrinsic value of work. Work makes the man, or the woman. Work qualifies us as valuable members of society. The jobless, the lazy, the indolent, are a burden to hard-working, family-raising taxpayers. But you can also read this story in another way. That work is unnatural.
Indeed, one would argue that once we managed to overcome our ignorance and superstition, and embraced science and technology, our prime goal has been to improve our lives. To sweat less. To live longer, but not in order to work more but to enjoy more. To return to that pre-farming state of existence where we do not need to worry about tomorrow, when we lived in the warm embrace of natural bliss.
We are the creators now and in come the robots, our creations, machines in our own image, imitating our arms and hands, building other machines for us, taking over the toil, and the need for us to carry out such heavy work. As the robots become more dexterous, and more intelligent – they start looking more and more like us. By way of extrapolation, one day soon, they may become capable of doing all the work for us. Shouldn’t we therefore welcome this? The end of work? The end of sweat? Unfortunately.. getting rid of work is complicated.
To illustrate the complication, I would like to tell you another apocryphal story that brings together two titans of the car industry, the industrialist Henry Ford II and the leader of the American Automobile Workers Union Walter Reuther. It is set the late 1950s. Ford is taking Reuther on a tour of a newly built highly automated factory, showing him the robots.
Walter – Ford says – how are you going to get those robots to pay for your union dues?
And Walter Reuther responds: Henry, how are you going to get those robots to buy your cars?
And that’s the Automation Paradox in a nutshell: if the robots do all the work, how will the workers – who will be jobless – earn money to buy what the robots produce? How will the jobless live in a fully automated world? The Automation Paradox informs the number one public fear of AI and robots, which is the loss of jobs. It also informs a slew of suggested policies to compensate for job automation. Universal Basic Income and Robot Tax are two examples. In fact South Korea has implemented a negative robot tax by reducing tax incentives for investment in automation. But, maybe, we need some fresh thinking about the challenge of the Fourth Industrial Revolution.
Maybe we need to think outside the box of economic scarcity, of winners and of losers, of haves and of have-nots. Because, for as long as the economics of scarcity persist, the robots will create more inequality, more scarcity, more misery. Maybe we need to think of the future economy not as one of scarcity but as one of abundance.
They say that futurists come in two flavours: prophets and wizards. Prophets warn of impending doom and ask to change our ways before is too late! Wizards see a way out of everything, by some kind of magic. I tend to side with the wizards. The magic out of joblessness and despair is to use our scientific knowledge and our technologies to radically change the economy. You see, technology is not a force of nature; technology is an ideology. And as such it is informed and guided by our values, beliefs, and aspirations.
We should therefore ask ourselves: to what ends shall we use automation? What purpose should automation serve? To enrich powerful corporations and the already super rich? Or to free humanity from having to sweat one’s face to eat bread?
How can democratize ownership of technology so that the bounty of automation is shared more equitably? How can we cut the Gordian knot of the Automation Paradox by moving beyond the dipole labor versus capital? What about knowledge? This year’s Nobel Prize in Economics laureate Paul Romer demonstrated how knowledge is an engine of growth. Can we spread the ownership of knowledge to the many? How can we ensure self-sovereignty of data – the fuel that runs in the algorithmic pipes of the Fourth Industrial Revolution -, quite possibly using cryptoeconomics and cryptogovernance?
Prophets warn of job apocalypse. Wizards conjure a green utopia of creative leisure for everyone. Up to you to decide whom to listen to.
Ernst Theodor Amadeus Hoffmann is considered the father of Gothic horror. They say that Mary Shelly wrote Frankenstein under the influence of having read one of Hoffmann’s stories.
My next story is a fusion of two of Hoffmann’s gothic tales, and goes by the title Coppelia, the girl with the enamel eyes. It is one of the most famous and widely performed ballets. The story involves an evil genius, Dr Coppelius, who has built a life-size, dancing, doll. His diabolical plan is to bring the doll to life using a human sacrifice. As it happens, he does not have to look too hard to find someone to send to the afterworld: a young man, by the name of Franz, gets smitten with Coppelia, the lifeless doll. He just can’t take his eyes off her.
However, this young man is already betrothed to a local girl named Swanildha, and Swanildha is simply not the kind of girl to sit quietly and let someone else snap her man away. After a series of mishaps and adventures she discovers Dr Coppelius’s evil plan, shows Franz his folly by dressing up as a doll, and saves her hapless fiancé from an untimely death at the hands of the diabolical inventor.
The Coppelia story is a wonderful interplay between the human and the human artefact, the mirroring of roles and identities between the real and the imagined. Humans falling in love with human-like artefacts goes back thousands of years. Ovid in the Metamorphoses, recants how the Cypriot sculptor Pygmalion fell in love with his creation, Galatea; and how he begged Aphrodite, the goddess of love, to bring his white marble statue to life. The goddess obliged, and Pygmalion and Galatea married and lived happily ever after, in what is probably the only Greek myth with a happy ending…
Stories such as these inspired the construction of mechanical automata in the 18th and 19th century, which in turn inspired Hoffman to write his story, which in turn inspired Charles Babbage to invent his Analytical Machine, the first general purpose computer. Stories make us dream, and then they make us want to realize our dreams.
How many of us have dreamt of the modern, cinematic, robot divas, in films like Metropolis, Blade Runner, and Ex Machina: Maria, Rachel, Ava.
The Coppellia myth lives on. Professor Hiroshi Ishiguro of Osaka University is creating life-like androids, mechanical imposters of real people. Staying in Japan, 2,000 “sex dolls” (not yet animated but very realistic looking) are sold in that country every year. They are called “Dutch Wives” and their buyers form bonds and relationships with them.
But, what is happening here? Can we really fall in love with “things”? With simulacra? With statues and androids and sex dolls?
Well, strangely enough, the answer is emphatically yes. We are victims of robot love because evolution has shaped our cognitive system to do exactly that! Let me explain myself. What makes us fall in love with robots, is exactly the same thing that makes us fall in love with each other – psychologists call it “theory of mind”. Our minds instinctively assume that other people have minds too, similar to ours. This cognitive ability emerges around the age of 3 to 4; and it is essential for empathy, for “feeling” for the other, for caring, for loving. Without theory of mind we would be a careless, selfish, species. We would never have literature, theatre, or art. We would be sociopaths.
Theory of mind compels us to project a “human-like mind” where there is none. If you have a dog, or a cat, or any pet, you know what I mean. We anthropomorphize animals all the time. But we also anthropomorphize inanimate objects, our cars for example (I often speak to my car, specially when it refuses to start on a cold day…). Religious veneration of rocks, statues and icons, are all examples of our minds projecting minds unto things.
In fact, you do not even need to see something in order to assume it has a mind. It is enough to listen to a human-like voice. That was the deep intuition of computer pioneer Alan Turning, when he suggested his “test” for telling if a machine is intelligent or not. If it sounds intelligent, it probably is.
Intelligent machines come also in disembodied forms, like Digital Assistants or chatbots. It is not necessary to think of intelligent machines as human-looking androids only. Think of Alexa, Siri, Cortana – the real digital divas.
And there are many different kinds of love as well, beyond the erotic, or the romantic. It is the love among friends, siblings, parents and children. So here’s the dilemma: as we increasingly forge relationships with artificial beings, can we trust them? Can they really be our friends? You see, a real friend is one that prefers to tell you the truth, however painful. So far, AI algorithms are trained to flatter us, manipulate us, nudge us to behave in a controlled way. Algorithms want to make robots out of us.
What if an algorithm told you that John will make a better husband that Tom? Would you believe it? Believe it blindly, without asking why? More and more people do exactly that. Over 61% of adults age 18-29 are using a dating app in the US, with 84% stating they do so in order to form a long-term romantic relationship. These dating apps are driven by AI algorithms – that tell you whom to love. How is this different from Dr Coppelius convincing Franz that he should abandon Swanilda in favor of a mechanical girl with enamel eyes?
Phillip K. Dick, the sci-fi novelist, possessed a deep intuition about a future where machines imitated humans: that is would be a future of paranoia. As digital avatars become more life-like and capable of resembling anyone, real or imaginary; as digital assistants can have conversations that feel completely natural, the distinction between real humans and artificial humans will begin to fade. It will not be too difficult for maleficent criminals to develop a digital avatar of you, and have that avatar empty your bank account, or commit a crime. Neither will be hard to imagine how the fake news of today will become the fake realities of tomorrow.
So let me suggest an addition to Asimov’s Three Laws of Robotics. Here’s my Fourth Law of Robotics: Robots must always declare they are robots and never pretend to be humans.
As for the laws of love, I would just leave them as they are…
The third story I want to tell you is from a book written some 2.5 thousand years ago by Plato, one of the founders of western philosophy. The title of the book is “Republic”.
For some context, Plato lived in very turbulent times. He was about 23 years old when Athens capitulated to Sparta, in the Peloponnesian War. The first thing that the victorious Spartans did was to replace Athenian democracy with the oligarchic rule of thirty aristocrats, known as the “Thirty Tyrants”.
Plato was pleased. He was an aristocrat, and one of the tyrants was an uncle of his from his mother’s side. Moreover, he was deeply disillusioned with democracy. (by way of note let me add here that Athenian democracy was a direct form of democracy, not the representative kind of modernity). The reason that Plato, and many other Athenians (mostly aristocrats) hated democracy was that they hated how easily demagogues swayed public opinion. Indeed, Plato felt that democracy was the reason that Athens had lost the war to Sparta. So he was pleased to see the end of it.
But what happened next troubled him even more. The Thirty Tyrants turned out to be even worse that the demagogues. They initiated a pogrom of their political opponents, and soon they begun to confiscate property belonging to other aristocrats. For the record, they were ousted by a popular uprising a year later, and democracy returned in Athens.
But for young Plato nothing was the same. Equally disillusioned both with democracy as well as oligarchy, he sought to imagine a better approach to government, and the Republic was the result of Plato’s thinking.
In the Republic Plato suggests that society should be ruled by a class of “philosopher-kings”. They are experts who possess the necessary knowledge for running a polis, a city, or a country – holding all three branches of government, the executive, judicial and legislative. Very importantly, the philosopher-kings hold very high moral standards. They are selfless. To make sure of that Plato suggests that they should not be allowed to have children, so that they are totally dedicated to the welfare of their fellow citizens and not bound by family ties and obligations. Philosopher kings would be selected from across the populace; among the brightest of children regardless of their family’s status or wealth.
Plato believed that the ideal society should be run by experts selected and groomed in a meritocratic way. In many ways, his argument sounds convincing. Wouldn’t you rather have experts running things than non-experts? Where the Republic fails is in attributing humans with unnatural and impossible levels of moral integrity and selflessness. History has shown that the first goal of all ruling elites is their self-preservation. Worryingly, by suggesting that society is separated in rulers and ruled, Plato’s Republic provides the philosophical foundation for totalitarianism, as the German philosopher Karl Popper has noted. For isn’t what communism, fascism and nazism suggest too? “Ideal societies” run by “ideal humans”?
But what if it was possible for technology to deliver omniscient experts who were also morally perfect? What if we had infallible AI’s instead of fallible humans running our countries? Could we see a new kind of totalitarianism rise in the 21st century, where the AI’s become Plato’s philosopher kings?
Making the right decisions in a complex, highly interconnected world requires superhuman abilities, let alone strong moral fiber. Big Data and AI will help us solve for global challenges such as climate change, environmental degradation, food security, health – to mention a few. Big Data and AI will help us make our cities better places to live in. The opportunity is indeed great.
But, as I said before, technology is an ideology. The same machines that can help us built smarter cities can be used to manipulate citizens and enforce obedience to the diktats of an elite. It is already happening in China, where personal data are collected and processed, and citizens are given a positive credit score for “good behaviour” and a negative credit score for “bad behaviour”. Totalitarian ideologies will use AI to solidify their rule. AI is ideal for citizen surveillance and control.
But what about democracies? How should we use AI to increase – not decrease – freedom and liberty?
There has been surprisingly little thought given to this question so far. Maybe because we take our democracies for granted. We should not. Also maybe because we have allowed ourselves to conflate freedom to choose with freedom to choose what to buy. The big tech corporations – that thrive because of our democratic institutions – are making incredible amounts of money by harnessing our data through AI to make better consumers out of us. Consumerism, however, is not democracy; it’s just consumerism.
In a Platonic Consumerist Republic the AIs decide what you read, what news you get, what you eat and where to eat, who you befriend and who you marry, how you entertain yourself. Because AIs know more about you than anyone, including yourself, they are much wiser than you. They simply know better. In this consumerist scenario AI is “augmenting” you because it enables you to make better choices, without the need to worry your head too much. You can just sit back, relax, and let the AI take care of everything. Just ask Alexa.
Augmentation sounds like a good deal, until you realise that every augmentation is also an amputation. Augmented by a car I do not need to use my legs anymore. Augmented by Alexa to keep my calendar in order I do not need to remember what I m supposed to do tomorrow. Just ask Alexa.
When AIs do everything for us, when we do not need to think, or take responsibility for our decisions, we become like children, we are infantilized. We live comfortable lives in apparent harmony, just doing what our smart overlords recommend we did. We become what Nietzsche described as the last men, and last women.
Ladies and gentlemen,
AI has been likened to electricity. And just like electricity AI is ushering a new industrial revolution, the Fourth Industrial Revolution. Electricity made all things come to life. AI is making all things intelligent. For an engineer this is an opportunity to design better systems to advance economic prosperity.
But something that will so profoundly change our world should not be viewed only through the lens of engineering and economics. Indeed, AI is a technology that needs to be deeply humanized – in order to be democratized – as soon as possible, to be transformed in our own image, lest it transforms us into something terrible. To humanize AI we need to come together across academic disciplines and walks of life and answer some age-old questions.
What does it mean to be human?
What should be our purpose?
What should we aspire to be in relation to ourselves, our families, our societies, our planet, the cosmos, the future?
These are questions that have tormented many big minds over many centuries, but in our present days of mindless entertainment and frivolous pastimes we seem to have become too cynical, or too lazy or lethargic to ask them.
It’s time to start asking those questions again, with renewed urgency.