Worshipping the AI God
#567
Last week the phone kept ringing about a Sothebys online auction in which a work by a humanoid robot sold for more than a million dollars – more precisely, for US$1,084,800 (AUD $1.66 million). The sale which ended on 7 November, saw 27 bidders competing for a portrait of mathematician, Alan Turing, called A.I. God. The artist was a robot named Ai-Da, developed by a scientist in Cambridge, UK.
Naturally the result went viral and everybody started asking that same, dreary question: “But is it art?” The one and only answer is: “Yes. Anything and everything can be art nowadays.” But it would make more sense to ask: “Is it good art? Does it move and inspire us?” This is where matters get tricky.
To sample the full flavour of this AI landmark, one needs to watch the online videos of Ai-Da talking about her work (I’ll respect her pronouns and say ‘her’, not ‘it’). Looking like a blank-faced 20-something with an Anna Wintour hairstyle, and shiny metal arms, Ai-Da talks in the manner of an eager young student explaining her work to the teacher. Unfortunately, her words don’t correspond to the mechanical movements of her mouth, which is creepy to say the least.
Ai-Da has created something that comes across as an excellent school project I’d score highly if I were her examiner. But she has succeeded on a purely formal, rational level, resulting in a soulless painting based on bits of old photos. If you expect a portrait to reveal something insightful about its subject, you need not bother with A.I.God. Our robot maestra chose Alan Turing because he played a big role in the early history of A.I. It’s like a child in the Young Archies competition submitting a picture of her dad. The complexity of Turing’s personality is not even hinted at in this fractured mask.
One feels sorry that Ai-Da, bright spark that she is, will never know the pleasure that comes from the happy accident, in which an artist sets out to do one thing and finds they’ve done something quite different – something they like better. The best part about human creativity is that it doesn’t proceed in a cold, logical manner. There are all sorts of subconscious influences that percolate to the surface, sudden distractions that rearrange an idea, moments of boredom and confusion that take one elsewhere, and so on.
Did Picasso plan Les Demoiselles d’Avignon and execute it precisely? No. It’s all over the place and much better for it. What Ai-Da lacks, and may never find, is inspiration - in all its many guises.
When AI can reproduce the countless random factors that influence human creativity, it will be time for artists to give the game away and become computer programmers. Even allowing for the technology’s capacity to keep learning and evolving at a staggering rate, I suspect it will be a very long time until it can match the human mind in terms of artistic creativity.
As I say this, I’m conscious there are many artists who are no better, and arguably much worse than Ai-Da, when it comes to inspiration. There’s a large, very successful sub-group of artists who simply decide on an image or an object and get it made in a factory to their specifications. Think of Jeff Koons or Damien Hirst or, closer to home, Patricia Piccinini. On the whole, there’s not a lot of difference between these programmatic art products and Ai-Da’s paintings. They are works of logical calculation, with themes and references mapped out to connect with contemporary preoccupations.
I daresay I’ve met artists, and certainly curators, with less human animation than Ai-Da. She may move her mouth in a strange way, but at least she seems friendly and enthusiastic.
There’s a second clichéd question that Ai-Da seems to encourage: “Is this the future of art?” The query seems predicated on the idea that art is always believed to be ‘progressing’ in some manner, but this is a dubious assumption. The myth of progress in art began to look threadbare as Modernism spluttered to a conclusion with Minimalism, then Conceptual Art. After the art object had been dematerialised, the only thing left to do was to bring it back, like a magician pulling a rabbit from a hat.
With Post-modernism, everything came back, helter-skelter, within inverted commas. When this orgy of irony had run its course, every movement and genre seemed to have returned. From this point artists were more likely to be judged as individuals, not as part of a school or movement. The only way the idea of “progress” could be kept alive, was via technology. The new avant-garde has embraced increasingly sophisticated forms of video and digital imaging, Virtual Reality, NFTs, the metaverse, and now AI.
Ai-Da, with her boring school project art, is the new cutting edge. As such, it was inevitable that buyers should have fought to acquire her work at auction. It’s widely believed the first example of any artistic product has an intrinsic value that can’t fail to escalate over time. Besides, there’s such an outrageous amount of money floating around at the top of the contemporary art market, that $1.66 million seems an almost modest sum.
Don’t forget that Beeple, the king of NFTs – but initially unknown in fine art circles – sold a work through Christies in 2021, for US$69 million (AUD $106 million). That kind of overnight success puts Ai-Da in the shade.
I’m not convinced that Ai-Da is as bold an experiment as the London Evening Standard’s attempt to use AI to revive their acerbic art critic, Brian Sewell, who died in 2015. In September, the Standard printed a review of the Van Gogh exhibition at the National Gallery, written by the AI Sewell. It was greeted with almost universal disdain – for the obvious reason that it sounded like a third-rate parody of Sewell – a writer who often sounded like a parody of himself.
Whatever one feels about Sewell’s trademark brand of invective, it was never as trite and superficial as the AI version. The first sentence, which tells us the Van Gogh show “is yet another exercise in sentimental hagiography,” sounds a false note that is echoed throughout the mercifully brief review. It makes one wonder whether Artificial Intelligence might not be a contradiction-in-terms.
If art critics are to be killed off, it won’t be AI that does it, more likely those media outlets that believe it’s best to run only ‘supportive’ articles about potential advertisers. As yet, the technology has not shown us it is able to find authentic artistic inspiration, or has any aptitude for capturing a writer’s distinctive tone of voice or thought patterns. Having spent so many years writing art reviews, I can’t imagine how a machine could handle an unusual association, twist a sentence, or find something original and to say about a work of art. If the Brian Sewell experiment is any indication, the technology is not even close. To make an impression, AI would have to overcome its inherent rationality, because one of the things that make us distinctively human is that we are not purely rational beings. How dull it would be, if all artists and writers were creatures of pure calculation.
Beyond the lifeless artworks and parodic reviews, there’s another concern about AI’s involvement in the arts: its complete amorality. When we consider the eruption of hate speech and disinformation released into social media by the simple instruction: “increase engagement”, one can see how an AI program could produce works of art and literature designed to tap into an audience’s fears and hatreds, producing negative propaganda with a skill that goes far beyond anything Dr. Goebbels might have attempted. The program would also learn from its successes and failures, refining its methods over time.
We’ve already seen what happened when AI was invited to produce ‘Norman Rockwell’ style pictures of America today. It unleashed a torrent of blatantly racist, anti-Semitic images, in which happy smiling people enacted unthinkable scenarios. All the algorithm was doing was drawing on sentiments that were already out there on the Internet. Given more precise and malign instructions, the potential for harm is unlimited.
We are irrational, emotional beings but AI is cold and logical. In any imagined confrontation it’s hysteria versus Realpolitik. What hope do we have? Very few people sit down and think through the root causes of their unhappiness. We want instant solutions and scapegoats. As the recent US election has demonstrated, nothing much has changed since Tom Lehrer wrote National Brotherhood Week in 1965 – nothing except the incredible technological advances that allow every hateful impulse to spread virally to all corners of society.
If every AI product were as benign as Ai-Da’s tribute to her daddy, Alan Turing, we might be able to view this high-tech turn as merely the latest contemporary art novelty. The worry is that beyond that stiff-faced dolly with the bobbed hair and robotic arms, there are AI superbrains busy rounding up all our demons so they can be organised into highly efficient regiments and sent out to do their worst.
The art column for The Nightly, is a less controversial proposition than last week. Cats and Dogs in Art & Design, is a shameless attempt by the National Gallery of Victoria, to reach a large popular audience. It draws on our attachment to our pets to bring together a lively selection of art and bric-à-bric from the permanent collection. It won’t change the world, but it’s a very appealing show.
The film being reviewed is Wicked, the Hollywood blockbuster based on the musical, based on the Wizard of Oz. If I say it’s not exactly to my taste, this is an understatement, but it’s fascinating to analyse the movie in relation to the current cultural and political climate in the United States. The mass cultural product is always revealing when one considers it has been tailor-made to a Hollywood studio’s idea of an audience. Universal is anticipating a gigantic box office success, giving the public the fairyfloss it needs to distract from the fact that they’ve just elected the Wizard of Oz into office.
PS. Here’s a link to an ABC radio interview on the AI question.