Working on the Blockchain Gang

Théâtre D’opéra Spatial: AI-generated artwork by Jason Allen

Some people imagine you can do anything with money. They are, obviously, people with very limited imaginations, but there are so many of them that they manage somehow to keep this absurd idea afloat on the ocean of popular culture.

In the last couple of years, these people have been getting very excited about NFTs – and if you don’t know that stands for ‘non-fungible tokens’, sir or madam, I envy you. A NFT is an image, sound or video file, created either by digitally photographing, recording or filming actual objects or sounds, or else by using one of the many computer programs available to create artificial ones. Whichever the method used, the ultimate product, in ultimate terms, is the same: a string of ones and zeros parked on a server somewhere in cyberspace, or encoded on a passive device like a CD-ROM.

NFTs are putative works of art. In reality, of course, they’re just data, some arrangement of bits and bytes that anyone can copy easily enough, even modify to their taste if they wish. But the original version, though humanly indistinguishable from any copy made of it, is uniquely identified by the data trail it has left behind in cyberspace since its ‘creator’ brought it into being. This purely theoretical uniqueness makes possible the existence of what you might call a digital-art market, which a lot of otherwise intelligent folk are getting awfully excited about.

NFTs appeal to two kinds of people. The first are those who collect art for its investment value, or for the bragging rights that come with ownership of a ‘masterwork’. They may or may not love the work itself; that is secondary. The most important thing about a work of art, for such people, is that they own it. These are the customers who will drive the NFT art market if and when it ever gets going.

The other people to whom NFTs appeal are commercial artists operating at a certain level. They may well be as capable of making ‘fine art’ as the poor fools who toil in hardship and obscurity, driven by their muse to create a thing nobody wants to buy. Their technical prowess may be as great or even greater. But these are bread-and-butter pragmatists who have learnt or come to believe that they do not possess the extraordinary ability (and luck) it takes to impress ordinary people who have never before seen anything like the picture, poem, concerto, movie or whatever it is that they have just produced: that they lack, I mean to say, the genius of a Greco, a Milton, a Rachmaninov or a Kubrick. Still, they know they have more than enough talent to produce something that, while not compellingly original, is still attractive enough to find a buyer. The customer might be someone who likes decoration for its own sake, perhaps, or wants something to hang in their offices or dental surgeries to soothe their clients, or to help sell their brand of vegan sausages or life insurance. Such buyers, too, form a market, even if it isn’t what we snobs normally call the ‘art market’.

Artists operating at this level – known as illustrators, graphic designers, visualizers, copywriters, music suppliers and a plethora of other names, depending on the media they work in – spend most of their time creating art that will pay the rent, put food on the table and so on, rather than the art that they would rather make. Some of them, if they grow prosperous enough to have spare time, will spend that time in self-expression, making art for art’s sake. Most of the time, the works so produced will exhibit no more aesthetic merit than the stuff they churn out during the rest of the week. This, then, is the type of artist – the type that nurses aspirations above his or her creative station – who is most attracted to the concept of NFT art and (in my experience) the first to try to ‘break into the market’ themselves. 

And this leads to a most intriguing paradox, because the type of artist to whom the idea of NFTs most appeals is exactly the kind of artist whose work does not command the kind of interest and reputation that (along with, of course, antiquity) make a work of art valuable. They – we, I suppose, and our name is Legion – are expert fabricators of kitsch, authors of pretty, briefly diverting trifles. Anyone who buys our NFTs is paying good money for something that nobody else is ever going to covet. NFT buyers, remember, aren’t interested in acquiring art for its artistic merit but only for the pecuniary and status value into which that merit can be converted. But if the work – the NFT – has no artistic merit, it has no pecuniary or status value either.

We seem to be in a Catch-22 situation here, folks.

§

To be honest, NFTs and their moneymaking potential fail to enthral me. Like ‘the cloud’, cyberspace, cryptocurrencies and the metaverse, they’re just another vaporous scam set up by crooks to part fools from their money by convincing them that something imaginary exists. Pull the other one, pal, is my usual reaction.

But it seems to me that the NFT delusion, peddled and bought into by the blockchain gang, is closely related to something else that’s been exercising the minds of those who make their living out of producing things that aren’t quite ‘art’ but share the same creative processes and aesthetic conventions. They’ve been palpitating about it a lot on social media lately. A few days ago one of my friends (a photographer) posted about some new ‘photographs’ – art images really – that had been created by an artificial intelligence (AI) program. OMG OMG, went the thread of his remarks, we photographers will soon be out of a job. Another photographer weighed in, pointing out that somebody had to take the original photos, so there would still be a human input. He didn’t explain why, in a world of spy satellites, automated webcams, sea turtles fitted with GoPros and robot telescopes photographing the early-morning light of Creation itself, the photos used by the AI still had to be ones taken by a human being. Another person (an architect) then came on the thread to warn, Jonah-like, that architecture and painting would be next to go. 

At this point I butted in to say that both these redoubts of human uniqueness had already fallen, only to be informed that I was wrong: AutoCAD and Adobe Illustrator and suchlike still require human input. It didn’t seem to occur to these guys that once you have AutoCAD and Illustrator and stuff like that, you’ve already transferred the creative process out of the real world and are working in an analogical space, manipulating ones and zeros rather than pens, brushes and paper. You’re not painting or taking photos, you’re doing mathematics. And how – you being a mathematical ignoramus – are you able to do that? It is always the machine that processes your ones and zeros into an image, or sound, or text. How long did you expect it to be before the human factor in that operation was eliminated altogether?

Well, now it has. But what is it, really, that the AI is producing in the form of these images, words, sounds et cetera? Is it art? And if it is, what makes it so? And then, of course, the Big Question: is it any good?

For the artists who make art for a living rather than under compulsion from their muse, these are haunting, terrifying questions. The obsession and terror come from the thought that – if AI productions can be art, and are good enough to pass for the ‘real’, i.e. human-made, thing – then we’re all going to be out of a job. And that, I think, is why this is the kind of artist who falls for the lure of NFTs. It sounds like a way to infuse value into work that may soon, otherwise, be worthless. Who is going to pay them for their art if an AI can create something equally acceptable for nothing, or next to nothing? All over the world, people who rely on their creative abilities to make a living are wondering how long they can keep on doing so. NFTs sound like one possible solution, and the concept has the art-market snob value of exclusivity built in, so it’s hardly a surprise that the idea tempts many commercial artists. But – as I pointed out earlier – they’d be foolish to fall for the scheme, because they don’t, by definition, have the talent to swing it.

§

On, then, to those terrifying questions about AI-produced ‘art’. 

Let us try to answer them.

The first two are pretty easy. Unless your definition of art is either proprietorial or snobbish – the kind of definition that excludes children’s drawings or images originally produced for commercial purposes – then yes, AI art is art. What makes it art? Well, what makes anything art? Here’s my working rule: any non-natural object designed to engage the emotions through the senses is art. This includes things like cars, shoes, even that hideous gold toilet-paper dispenser on Putin’s yacht. A thing doesn’t have to be good, nor be useless for any practical purpose, nor command fourteen times its reserve price at Sotheby’s, to make it art; if it just makes you want to look at it, or listen to it, or touch it, or whatever, that will do. It’s art.

This, then, brings up the valid but dicey question of intention. The human artist aims to convey some thought or feeling. The AI, as far as we can tell, is just following instructions. Does that make a difference? But then, the effect of an artwork on its audience is rarely exactly the one its maker intended. Does that make a difference?

Stalemate.

So let’s leave it and move on to the big question. Is art produced by AIs any good?

In my opinion, it varies. Recently, I was sent a New Yorker article about an AI that can write poetry in the style of any famous poet. All you have to do is feed the beast a few examples of the poet’s oeuvre and it will produce reams and reams of the Good Stuff. The writer of the article seemed to think it was pretty good. He included a couple of examples of poetry the AI had written in the style of Philip Larkin. Here’s one.

The Invention

Money is a thing you earn by the sweat of your brow

And that’s how it should be.

Or you can steal it, and go to jail;

Or inherit it, and be set for life;

Or win it on the pools, which is luck;

Or marry it, which is what I did.

And that is how it should be, too.

But now this idea’s come up

Of inventing money, just like that.

I ask you, is nothing sacred?

Now, this isn’t gibberish; far from it. And it contains words that you can imagine Philip Larkin using, and captures some of the oddities of his syntax and punctuation, so good work on that. But the idea it expresses is mundane and trivial. The lines evoke no feeling and certainly none of that throat-catching emotion that grips you when you read a line of verse that the poet has got exactly right. And – fucking hell, as Larkin himself might have said – it doesn’t even rhyme. Larkin only wrote about seven or eight poems, out of the hundreds he published, in free verse; and he was a rhymer with a particular style, too. But even if we give the industrious AI a pass on that, The Invention could never be mistaken for anything by Philip Larkin. It can’t even be mistaken for a poem.

I thought I would demonstrate the difference by reproducing an actual Larkin poem (that doesn’t rhyme). I reckon a good choice is Days: far from his best, but about the same length as the impostor. Hope no-one sues me...

Days

What are days for?

Days are where we live.

They come, they wake us

Time and time over.

They are to be happy in:

Where can we live but days?

Ah, solving that question

Brings the priest and the doctor

In their long coats

Running over the fields.

If you can’t see any real difference between that and The Invention, gentle reader, maybe NFTs really are for you after all.

But there are other AI-produced works that are a lot more like art in that they evoke, through the senses, real ideas and emotions. Some time ago there appeared, in my Facebook feed, an album of ‘photos’ of what looked like a Gothic horror-fantasy-science-fiction story come to life. Many of the people commenting on the photos weren’t sure whether the images were real or created by some kind of photo-manipulation; quite a few of the commenters seemed convinced of the former. And the images are, indeed, shockingly ‘realistic’, though they depict things that certainly do not exist on Earth. 

The album, which you can view in its original form as an imgur.com portfolio, is called The Curious Journeys of A.I. Midleton and it was produced (as the title hints) by a AI program, called Dall-E-2. It works rather like the poetry AI does: you feed it with words; it uses them to search the internet for image files, which it then combines and manipulates to create ‘art’.

Here, though, there was a difference. The person interacting with the AI didn’t just settle for the first images they got; they modified the words, picked out related images, modified again and selected again, and when they’d got something they were reasonably happy with, they worked on it in PhotoShop (or something like that) to make them even more convincing. So this was, if you like, a human-AI collaboration. If it’s art, it’s art that a human being was closely involved in making.

But then, just a few days ago, I learnt from CNN that in faraway Vail, Colorado, an AI-produced work has won first prize at a state art fair. Sadly, the work was not entered by the AI that created it (real world, you see; difficult) but by a computer-game designer named Jason Allen. He used another open-source AI, Midjourney, to ‘commission’ it from. Apparently he spent eighty hours in consultation with the AI over it, but I don’t think he used any image-manipulation software like Adobe Illustrator or PhotoShop. I’ve fooled about with Midjourney myself, just to see what the fuss was about; it seems to be a bit more ‘realistic’ than Dall-E-2 and to require less post-processing. Anyway, the result is a pretty convincing ‘artwork’ that you, if you didn’t know better, would certainly believe was created by a human being.

The controversy over this little affair is already pretty hot. Some people say Allen cheated, or that the picture isn’t art. I don’t know about that; it looks like a picture to me, and the sight of it certainly evokes some kind of feeling, though – for me, at least – no very strong one. But it seems the whole argument is a bit cockeyed, because, once again, the human involved in the creation of the piece was picking and choosing, and combining, what Midjourney offered him until he got the result he wanted. 

So whether it was AI-generated or not, and whether it was art or not, it was finally a human being who decided how the picture should look. Shouldn’t that be enough to quash the controversy? The AI might have created and presented the original images, but the final result was completely dependent on human judgement.

Let’s take that argument one step further. A work of art is never the sole creation of the artist; it is those who experience it through their senses who ultimately dictate whether it is, or is not, to be regarded as art. So even if a one hundred per cent AI-created work does go on exhibition and win first prize at an American country fair, or even fetches a million pounds at Christie’s, there is still an enormous human input involved: that of the viewers, the listeners, the audience. As long as they accept the thing as art, it is art. 

§

Which brings us back to the very understandable anxiety of those of us who produce ‘creative work’ for a living. If these pesky machines get any smarter, we’ll starve.

Well, you know what. The market for work by commercial artists, writers, composers, filmmakers et al has been shrinking in monetary terms for a long time now. The average value of work (measured by what the creator can charge for it) has crashed. The internet certainly has had a lot to do with this; by removing barriers to entry, it has greatly increased the number of people competing for work in the commercial arts, while at the same time generating business-model changes and economies of scale that have put traditional middlemen and gatekeepers – agents, publishers, A&R departments in the music industry – out of work, too. I see no possibility, for a decade or so at least, of this situation changing. The internet, after all, is here to stay, and with it e-commerce and a legion of other inhuman, soul-destroying innovations. Tough.

I think what we are really seeing is the death of the mass market for human-made art. Art-for-a-purpose – advertising, interior decoration, popular entertainment – will henceforth be produced by machines and the human race, as a whole, may well be none the worse for it. But what about the artists? What will happen to them? 

Perhaps they will find themselves in a situation much like that experienced by practitioners of the arts in the ages before mass media and consumerism came to stay. They will, once again, look to individual rather than corporate patronage or the big money of advertising to sustain them. Their egos and the dignity they stand upon will undergo a salutary deflation as they find themselves reduced, once again, to the status of street performers, courtiers or servants. The vast sums a few big names now command will no longer be paid. The absurd, starvation-level fees upon which many commercial artists and writers subsist will cease to be paid at all. Anyone who is not an artist or author by compulsion will stop trying and go and get a boring office or service-industry job instead. In sum, the population of those who think of themselves as artists will fall to maybe a thousandth of what it is now; the market for human-made fine art will shrink to perhaps a hundredth of its present size; and the average income of those who are still in the game will rise, perhaps, tenfold from what it was before the rise of Amazon and its fellow cyberslavedrivers. A lot of us, artists and non-artists alike, may finally come to look upon this as a good thing. The arts, which are in a state of creative prostration for the most part, may even experience a renaissance.

Or maybe not. Predicting the future is a risky business (something the NFT advocates would do well to remember). As for the rights and wrongs of the case, I abhor making ethical judgements about aesthetic matters. I have my tastes and prejudices, but I don’t really feel the need to explain or justify them to anyone. But then, I am lucky enough to be, for the most part, retired from the world of creative commerce; I have less skin in the game than most of the poor guys and gals who are worrying about the ‘AI takeover’ on social media. My advice to them would be to make up their minds: either follow your own muse and prepare to starve for it as van Gogh did, or give up being ‘creative’ and get whatever job you can – and that you can stand. But whatever you do, don’t fall for the NFT scam, because it combines the worst of both worlds: no money (once the pecuniary worthlessness of NFTs becomes common knowledge) and no artistic satisfaction or fame for you, either.