In principle, because the brain obeys the laws of physics, computers can do anything the brain can do.
—Murray Shanahan1
The brain is just a computer like any other. … Traits previously considered innate to humans—imagination, creativity, even consciousness—may be just the equivalent of software programs.
—Demis Hassabis2
The time when humans can have meaningful conversation with an AI has always seemed far off and the stuff of science fiction. But for Go players that day is here.
—Andy Okun3
Will computers ever think like us? Could they ever have flashes of inspiration like we do or come up with mad ideas? Could they invent something no one ever thought of before and never thought was needed? Could they dream up the plays of Shakespeare?
Or do they need to? Perhaps they will function in totally other ways than human beings, come up with ideas just as great or solutions just as effective but different from the ones we would come up with.
This book strives to look into such questions. It is about creativity in the age of machines—our creativity and their creativity.
What do we mean when we talk about being creative? Is it that flash of inspiration when we suddenly think up something new—a melody, a poem—that no one has ever thought of before? That eureka moment, like Archimedes in the bath, when we suddenly see the solution to a knotty problem with blinding clarity? Is it the creating of something from nothing, like a novelist dreaming up a story or some outrageous science fiction scenario? Is this what it means for human beings to be creative?
Today we are rapidly developing ever more sophisticated machines, computers that can think. Perhaps one day soon machines will be able to think like us. Or will they evolve in a totally different direction? Will they think differently from us, function differently from us? Will computers too be able to be creative? In this day and age, we are going to have to rethink what we mean by thinking and what we mean by creativity.
Imagine a world in which machines produce bold new works of art and music, discover scientific theories, write stories, make business decisions, tell jokes, or function as the brain that operates a robot. Imagine an AI as the next James Dyson, dreaming up a whole new generation of vacuum cleaners. What will this mean for the future of our jobs? Perhaps the next Turner Prize winner might be a machine, or the next Pulitzer Prize winner, or the next Frank Gehry, whose creative projects already rely a great deal on computing. Imagine if the next Keats was a computer, or the next Beethoven. Suppose the next legendary chef was a computer, or the next stand-up comic. In the future, might your lover or your new best friend—or your therapist—be a robot? And what place will humans occupy in this brave new world?
Can humans and robots exist side by side? Will machines’ intelligence surpass ours? And what will this mean for the future of the human race?
To find answers to these and many other questions, I’ve gone to the experts. I’ve interviewed scientists working on frontier problems in computer thinking to find out what the latest developments are, what might be possible in future, and what problems could arise. I have focused on key players who are developing AI that creates art, literature, and music. All of them are doing extraordinary work, pushing forward the frontiers. And all have insights to impart.
Tony Veale of University College Dublin claims that “the question of whether computers are truly creative will fade away as people value the results.”4 Ian Goodfellow of Google, inventor of the groundbreaking generative adversarial networks (GANs), asserts that “machines are already creative.”5 Gerfried Stocker of Ars Electronica suggests that “we might need machines to create art as a way of communicating with them and understanding them.”6 But Allison Parish of NYU demurs: “I’m going to be a hardliner and say that computers cannot be creative.”7 Hod Lipson of Columbia University says that “the more intelligent AI becomes, the more sophisticated its art will be.”8 Kevin Warwick of Coventry University studies cyborgs. He says, “Creativity in machines is there, whether we humans understand it or not. If we can figure out how memory works, then chips will go beyond memory, and will improve our creativity enormously.”9 Douglas Eck of Google tells me, “Technology is special because it gives us AI, and AI may create things so beautiful that we may care differently about them.”10 Murray Shanahan of Imperial College in London was the technical adviser on the film Ex Machina. He states emphatically, “In principle, because the brain obeys the laws of physics, computers can do anything the brain can do.”11 Blaise Agüera y Arcas of Google says provocatively, “When we do art with machines I don’t think there is a very strict boundary between what is human and what is machine.”12
* * *
Very often, the greatest discoveries are of things we never realized we needed. It’s not so much a question of solving a problem as of seeing a problem that no one else had even known was there. Take Chester Carlson, for example.
Chester Carlson was an affable-looking thirty-two-year-old with a degree in physics and a restless imagination. He had held a series of low-level scientific jobs, but he was bored. So he decided to try law school. But he was so broke that he couldn’t afford books and spent his time sitting in libraries, copying passages of text. “There should be some way of making copies,” he thought. There were duplicating machines, but that involved making a master copy first, a complex and time-consuming process. Working in his tiny kitchen in New York City, in 1938, Carlson built a device that could copy a document onto a second piece of paper with no intermediate steps. The process involved electrostatics and powders that gave off noxious fumes. His device was primitive, and experimentation was dangerous. If only he had the funding, he was sure his invention would take off. But by this time, he had used up his savings and his clothes were threadbare.
Over the next eight years, Carlson tried selling his primitive photocopier to over twenty companies, including IBM. They all turned him down. None of them could see any point in the concept. After all, who would want to make thousands of copies of anything? In addition, the executives he pitched to were put off by someone who looked more like a mad scientist than a captain of industry. As Carlson later recalled, they couldn’t see the point of disseminating information, of spreading around the contents of memos and reports.
Finally in 1946, John Dessauer, chief of research at the Haloid Company in Rochester, New York, a fledgling business that produced photographic paper, heard of Carlson’s invention and decided to take a gamble. By 1956 Haloid, with Carlson on board, had produced the famous prototype 914 Paper Copier. Some billions of dollars later, the company changed its name to Xerox Corporation. Nobody had ever needed a copier or even imagined one until Xerox produced one.
A little over a decade later, Xerox set up its own version of Bell Labs, AT&T’s think tank. It was called Xerox PARC (Palo Alto Research Center). One of its first inventions was the laser printer. But by then Xerox was after bigger game than photocopiers: the computer revolution had begun.
About this time, a young man named Alan C. Kay arrived for an interview. His interest in computers had been sparked by a stint in the air force, where he programmed the huge IBM 1401. Afterward, Kay obtained a PhD in computer graphics, in the process learning about systems that allowed computers and humans to communicate by moving icons on a screen.
In the course of his interview at Xerox PARC, Kay was asked about his dream project. “A personal computer,” he replied.13 His interviewer was incredulous. This was the age of mainframe computers that filled a large room. What else could one possibly desire? Draw one, he challenged Kay. The resulting sketch looked much like today’s laptops. “Yeah, right,” the interviewer replied (in other words, “Who needs this?”), just as IBM had replied to Carlson.14 A few years later, Bill Gates and Steve Jobs took up Kay’s idea, and soon everyone needed their own personal computer—their own PC. Kay’s invention had quite literally changed the world and changed our concept of what we needed.
Kay’s brilliant realization was that trying to guess the future was a waste of time. “The best way to predict the future is to invent it,” he said.15 By the time the twenty-first century rolled around, this mantra had changed the very concept of creativity.
Creativity is the last great mystery. Where do ideas come from? How does a new idea pop into our minds? How can we cultivate it? How can we nurture and encourage our own creativity? I’ve worked on the study of creativity for decades and have come to the conclusion that it has certain qualities, certain characteristics shared by all thinkers across diverse fields.
Have you ever noticed that the keypad on your phone is laid out in a very specific design—three rows of three numbers starting with 1, ending in 9, and a fourth with a zero in the middle? (On today’s twelve-button keypad, the 0 is flanked by star and pound keys.) Has it ever occurred to you that this configuration might have anything in common with Einstein’s discovery of relativity theory?
The keypad was the brainchild of John E. Karlin, at Bell Labs, who applied insights from the behavioral sciences to telephone design. At the time, no one had ever thought there was any connection between the two. For Karlin, the dynamics of using a telephone required far more than speaking and hearing. Karlin was a man of many dimensions. In college in South Africa, he cast his net wide, studying philosophy, psychology, and music; he was a violinist in the Cape Town Symphony Orchestra. He then completed a PhD in mathematical psychology at the University of Chicago and went on to study electrical engineering at the Massachusetts Institute of Technology (MIT). He joined Bell Labs in 1945, where he made critical design changes to old-fashioned rotary telephones. He was also instrumental in changing phone numbers to include only numerals, not letters; the large number of telephones coming into use made this change essential.
By the late 1950s, rotary dialing was about to be replaced with touch-tone dialing, touted to be much faster. The problem was what keyboard design—that is, keypad design—would be best for speed and accuracy. Several were considered. Based on empirical research on how people use phones, Karlin’s turned out to be the most intuitive, the most user-friendly, and hence the quickest way to make a call. It has been with us ever since.
For Karlin, it was a creative leap. Similarly, Einstein hit on a hitherto unrecognized connection between the theory of heat, thermodynamics, and the motion of bodies in space and time. The result was his theory of relativity.
Both these leaps involved someone realizing the relationship between areas that at first sight seemed to have nothing to do with one another.
* * *
But could a computer spot these connections and come up with a better keypad design or even a new theory of relativity or a new style in art? Can computers be creative? And if they do start to think, need they think in the same way that we do? Perhaps they will be creative in entirely different ways from us.
Computers are already trespassing on territory traditionally considered to belong only to very brilliant and perhaps rather eccentric individuals. Machines have defeated champions at complex games such as chess, Jeopardy!, and Go. They can recognize faces, generate dramatically new and unexpected forms of art and music. Google’s Project Magenta aims to have its computer produce compelling art and music without being preprogrammed so that eventually it will be taken seriously as a creative artist. Computers are already being used to develop driverless cars and are having an impact on law and medicine. Today everyone lives surrounded by devices—phones, cars, remote-controlled ovens, heating systems, all connected to the web, via the so-called internet of things. All this is underpinned by computer systems performing tasks that we have always assumed require human intelligence. This is what we call artificial intelligence (AI).
So what is machine creativity? Does it or will it differ from human creativity? If so, in what ways? Will machines ever have consciousness, a trait usually assumed to go hand in hand with creativity? Will computers come to have emotions? Could a computer flying a plane experience an emergency situation and feel fear, bringing about an “adrenaline rush” and providing that extra impetus to think outside the box and avert disaster? Would a robot ever fix its hair in the mirror, as Arnold Schwarzenegger does in Terminator? Could machines suffer, and have awareness of themselves and others, like the replicant in Blade Runner?
* * *
We humans are slowly and imperceptibly merging with machines. Think of the intimate relationship you have with your cell phone, which is actually a tiny computer. Electronic implants are no longer science fiction. There is already a cochlear implant, an electronic device that replaces the function of the damaged inner ear, although this is not connected to our brain. In twenty years, we may have chips implanted in our brains that will give us direct access to the web and thus to all known knowledge and enable us to use that knowledge to create new knowledge. When brain implants become reality, we will no longer ever need to forget a name or face. Scientists with devices in their brains will be able to scan thousands of pages of research material in seconds and use it to make and weigh hypotheses.
There is already at least one computer that can do just this: IBM Watson, which became grand champion of the complex quiz game Jeopardy! Its victory was so impressive that for the first time people began to wonder seriously whether it had actually been thinking. As David Ferrucci, team leader of the Watson project, said, in an oft-repeated quote, “Can a submarine swim?”16 From which we can extrapolate the answer: “Yes. But not like a fish. Better.” Similarly, when chess grand master Garry Kasparov played against the computer Deep Blue, he reported that he sensed “a new kind of intelligence.” Perhaps soon there will be no limits to our creativity—or to that of machines, either.
People often say that computers cannot be creative because they merely follow the instructions that we, their human creators, have embedded in their algorithms. But surely this is akin to saying that Mozart couldn’t have been creative because his father taught him music and should therefore have all the credit for his son’s achievements. Today’s computers often transcend their algorithms to create new forms of art, literature, and music, just as we transcend what we have learned. This is what we call creativity.
But before we can get to grips with the creativity of machines, we first need to understand exactly what human creativity is and means.
* * *
I will start by focusing on the great thinkers and creators who have radically changed our view of the world. These are people whose extraordinary powers distinguish them from most others and cannot be attributed just to hard work. We can call them geniuses. Although we can probably never equal them, we can learn from and be inspired by their thought processes, by looking into the ways they work. If we can understand what highly creative people and geniuses have in common, that may enable us to increase our own creativity. I have devoted decades to this study and have identified hallmarks of high creativity and of genius, which we will examine in the next chapters.
The great question is whether computers too can develop these qualities, or whether they will develop their own ways of thinking and begin to operate and function autonomously—not as replica people but as an altogether different and independent form of intelligence.
This will involve looking into the “lives” of computers, exploring their creativity, their innermost thoughts, to what extent they may be similar to ours and to what extent different. To do so, I will look in depth at the extraordinary new art, literature, and music that computers are creating today.
In the end, we will probably never be able to grasp how machines understand purely on the basis of how we understand. Writers on AI frequently offer a dystopian view of the future, a view of a world ruled by hostile machines. But this need not be the case. AI may well not be the final link in a chain of technological innovation stretching back to the steam engine and beyond, one engine following another until human beings disappear, to be replaced by machines.
There may well be a happier ending to our story.