Recently, I played a series of symphonic movements for a class. Some were by Mozart, and others by other composers. With a little practice and guidance, the class picked up a rough impression of Mozart’s style, as distinct from the other works. The last piece I played was by David Cope’s software Experiments in Musical Intelligence (better known as EMI or Emmy). Emmy was designed to emulate other composers’ styles as closely as possible, and I wanted to test its effect on a class that wasn’t aware any of the music they would be hearing was written by software.
Of course, my real purpose was to test their reactions to algorithmic composition in general. One student, who’s preparing for her final school exams, gave a comment that’s been fairly exemplary of those I’ve heard when I bring up the topic: “You want to know that there’s a person writing the music. Otherwise how can it be special?”
Cope mothballed Emmy in 2003, and has channelled much of his subsequent work into another algorithmic composition project, Emily Howell, which uses outputs from Emmy and Cope’s training with an association network to generate music in its own style. It was only when I played Emily Howell’s music for the class that that same student was taken aback. She knew the piece. It’s in her study playlist.
The idea of algorithms that create art, and that create music specifically, is fascinating, and the more research I did on it the more interesting I found it. I think this piece should be a good primer for anyone who’s interested in the topic.
You may be quick to identify a portrait unveiled this week in Amsterdam as a never-before-seen painting by Rembrandt. With a calm gaze, mouth slightly parted, and wearing a frilled collar with a wide-brimmed hat, the man resembles the sitters the Dutch painter so frequently depicted. Rather than dabs of paint, however, this portrait consists of pixels—148 million of them, to be exact, all created by machines and captured in a 3D-printed painting.
We’re living in a world where computers can make convincing original works in the style of Rembrandt and Mozart. That is amazing.
I’m not convinced that it means artists’ jobs are under threat from automation in the same way as people working in other fields. Mozart and Rembrandt have been dead for centuries, and while their music and art still draws people to concert halls and museums, no composer or artist has ever gained success by writing or painting in their styles. Their originality is a key part of their popularity.
Computers now can create millions of works that are functionally indistinct from those by famous historical artists, and I doubt we’re that far from having them create entirely original works. But the function of art is more than the beauty of the melody or the control of the brush stroke; it’s the connection it makes between creator and audience, and that connection can’t exist, I believe, between human and machine.
It’s not the first time something like this has happened. In the nineteenth century, the falling cost and visual fealty of photography meant that traditional visual art didn’t need to recreate the physical world; doing so was redundant since photographs would do it more efficiently and more faithfully. So visual art started to move away from more realistic depictions of the world to more abstract ones, creating works that only humans could.
One of his most popular (and best) is Humans Need Not Apply, a chilling look at the future of automation and industry, and the potential massive spike in unemployment that may come about in the next fifty years as computers become better at human jobs than humans are. It’s about fifteen minutes long, and worth watching.
While I can’t dispute many of the facts, I think the video gets it very wrong on some aspects of creative professionalism. Grey is a utilitarian. He openly admits on his podcast, Cortex, that, for him, music is a tool that helps him get work done. Elsewhere, he’s sardonically mocking of poets and artists (as he is in the video above). That’s all well and good. Not everyone can be an art lover, and not everyone should be. But a utilitarian view of the arts leads to a fairly simple notion of art as something trying to accomplish a job.
In the last two weeks, music has lost two titans of two genres: the austere, difficult, polarising composer (and undeniably brilliant conductor) Pierre Boulez, and the chameleonic, gloriously weird rock1 musician David Bowie. The two men had little in common, and as far as I knew never crossed paths except in passing. But their music mattered to people, and it mattered far less because of its qualities than because those qualities were the result of human work.
David Bowie’s early rise to fame came about because he revelled in his weirdness. He was, as Hilton Als noted in his New Yorker piece, “that outsider who made different kids feel like dancing in that difference”. No computer, no matter how good the music it made, could ever forge that connection with people. A glance at Twitter over the past week shows how keenly his loss is felt as a personal one by people around the world.
By contrast, Hatsune Miku, a fully computer-generated performer from Japan (if you haven’t heard of her, then yes, really), certainly has her core group of fans. And I have no doubt that her fans enjoy her performances, and listening to her singing. But were she to vanish tomorrow, through some freak accident of data loss, all her fans would really lose would be her songs. With Bowie, they lost an icon.
Pierre Boulez’s early music became famous (and notorious) for its high degree of mathematical precision—not only the notes, but the dynamics, tempi, articulations were all rigorously figured out beforehand. Regardless of the emotional content of the music (which some listeners passionately defend), that is a type of music that is surely highly suited to a computer’s labour. But again, part of the appeal of Boulez as a musician is the fact that a human could create and hold music of this scale and complexity in his mind. Who would be interested if it was automatic? If his music had been made by a computer—a machine which necessarily would have found it easier—it would have been less interesting.
Recently, the American composer Andrew Norman made exactly this point on performing. From a New York Times interview by Will Robin:
“By thinking of the orchestra as only a sound-making machine, we’ve actually eliminated a huge part of what makes a concert experience amazing,” Mr. Norman said. A laptop, he pointed out, easily supersedes what the symphony can offer in terms of sonic power and flexibility. “What makes an orchestra special, for me, is not actually the sounds that it makes but the fact that there are a hundred human beings doing that, right in front of me,” he added. “In a way, it’s performance art.”
There are already computers that can generate music; even ones that can do a job of imitating Mozart well enough to fool Mozart experts. But short-term existential crises aside, these works become curios—interesting for how they were composed, but passed over because there’s nothing to get your teeth into. No composer weeping over the streets at the beauty of the sound of a fire sergeant’s funeral. No artist in a fit of horror at war dedicating himself to representing that horror in black-and-white. No writer so full of self-loathing that he imagines himself transforming overnight into vermin.
That’s not to say there won’t be room (not to say a market) for computer-generated art, but only in its most functional sense. What if I told you your film could have music by John Williams? Would you say no? What if the other option was Beethoven? Or night clubs, whose music—beyond the compulsion to follow certain trends of fashion—is essentially background noise for socialising.
But a computer-generated Beethoven symphony? Would we really want that? Beethoven was great, and we have piles of his music to listen to now—but he’s been and gone. Music has changed since his time, and to go back is pointless. And any originality expressed by a computer is uninteresting, not in spite of its lack of imagination, but precisely because its imagination is theoretically limitless.
Maybe I’m wrong. Maybe next generation’s Hatsune Miku will be one with a personality and life programmed by her team (as authors program characters—no judgement here), able to inhabit independently a virtual world. People can definitely come to love fictional characters, and maybe they’ll love her as they love Katniss Everdeen. Maybe the following generation will see a wholly computer generated performer, with appearance, personality, life all created by computers with the last human interaction having occurred thirty years before.
But my suspicion is that for the people who find human connection in art, art made by humans will always be essential. At least until computers can imitate a full, creative human mind. And then we’ll have plenty of new philosophical issues to deal with.
Last week, on the podcast Exponent (and to a lesser extent, in this blog post), Ben Thompson detailed a political position I hadn’t been familiar with before, but one which I find interesting. In short, as technology has a greater and greater impact on society, and as its presence costs more and more people their jobs, it is in technology companies’ best interests to lobby not only for less regulation (the clarion call of so much business), but also for higher taxes and an assured “ground floor”, economically, so that those people whose jobs are lost through technological disruption are not left with nothing. This way, the people whom regulations are supposed to protect are protected by the safety net, and the corporations have the freedom to grow as they please.
The reasoning is this: a job done by a computer is not of net benefit to society until the person whose job was lost is contributing something new. Otherwise, nothing is gained overall. The technology companies benefit through having fewer restraints on the ways they can develop; those restraints are less necessary if people are assured a stable means of livelihood anyway; and that better quality of life can be achieved through revenue generated from higher taxes. In both the businesses’ and the governments’ cases, something is given and something is gained.
To be sure, it’s not a flawless plan. I’m sure that one of the common criticisms levelled at it will be that if you pay people for doing nothing, then people won’t do anything. I’m more optimistic than that. I think, left without the constant worry about meeting basic needs, most people will try to occupy their time in ways that fulfil them, whether that’s making art, or starting a business, or whatever.
In any case, we may know soon enough whether a basic wage for everyone can have positive effects on a society: the current government of Finland are in the early stages of carrying out the experiment.
Equally, those more skeptical of the corporate world would argue that no corporation would lobby for higher taxes—these are self-interested entities, after all. But they wouldn’t be doing it out of the goodness of their hearts, but as a trade-off for reduced regulation (which, far more than tax, is the bugbear of many large businesses, especially tech companies). If the total cost of a computer doing a job plus a higher general tax rate is still cheaper than an employee (and it could well be), then the company still makes a saving, and the sooner that change is made, the more money saved.
More concerning to me is the ability of small and of poor countries to fund a project like this. There’s a compelling argument to be made that the tech companies who gain the most customers now, firstly, will become far bigger than any company that’s existed so far in human history, and secondly, will continue to dominate for the foreseeable future. Services become verbs—Google, Skype, Uber—and these services, as U.S.-based entities, will pay the majority of their taxes to the U.S. government. Good for American citizens.
But these services have a much smaller stake in other markets they choose to enter, and both the incentive and leverage to push local taxes as low as they can. Rich citizens (or visitors) in these countries demand the same services they get elsewhere in the world, but the companies that make those services can choose whether to enter the markets on their terms. They may be amenable to high tax at home, but would they be so willing to pay it everywhere else too?
These smaller countries, and these poorer countries, would be thus less able to compete with the rest of the world, and so would slip further and further behind in economic stakes. As a result, their governments would become less and less able to support their citizens under any system. In the Internet’s winner-take-all economics, maybe somebody still has to lose.
As I said, though, I’m an optimist. I think that most national systems will tend towards this sort of arrangement sooner or later, as a consequence of Internet economics and the needs that will arise because of it, in the same way that most western countries now have some form of socialism. (Not enough for some; too much for others, I know, but such is politics.) If companies can be compelled to pay high taxes in all territories where they operate—if that becomes the norm—then the revenue generated from that can keep people fed when their jobs disappear. Growth shoots already exist; the only real question is when it happens—and I believe the longer it takes, the worse off everyone will be.
As for whether people will still have things to make and sell, no matter how good computers get at making things (art, design, music, yes, but also sofas, clocks, kitchenware), I think people will hold onto a romantic attachment to the human-made. The down-side of Internet economics is that it allows companies to become enormous to a degree previously unthinkable, but it also—assuming it remains open and free as it should be (vote wisely, folks)—allows anyone with a computer and an Internet connection access to the largest market in human history.
This is Grey’s biggest error in judging the creative professions: he assumes that it’s a popularity contest. It isn’t. In order to get by, a creative person needs only a few thousand fans. Amongst the billions of people connected to the Internet, that is a fraction of a fraction of a percentage. Moreover, there’s a tangible thrill for fans in discovering something that nobody else knows.
For an artist in the twenty-first century, finding fans is hard, no doubt. But it’s far easier than it ever has been before. And Grey ought to know this. He’s done it himself.