I imagine an early maker of figurative art, long gone now, maybe forty thousand years gone. With startling economy, a few curves and blocks of colour, they create an image on a wall: an ox, a pig, a deer, a rhinoceros. They have seen the animal. They imagine what they have seen – literally, in the primary sense: they hold an image in their mind. They reproduce that image on the wall, with a finger, a stick, earth, blood, spit, vegetable matter. They know that the image is not the real animal. This is not the creature. But this is what I saw.
They call to one of their companions. Come and see. I cannot know how the two of them communicated – what language, or ancestor of language, they used. But the artist shows the image to the observer, and the observer now holds an image in their mind. I recognise this. This is a creature I saw. They imagine something like the creature the artist imagined.
A thought, a concept that had existed only in the artist’s mind, has leapt into the mind of the observer, through nothing more than the daubing of pigment upon a surface. It is wondrous, a kind of magic. It is, perhaps, a source of joy: The world you see seems to be the world I see; I begin to understand you; now we need not feel as alone.
Born into an age of complex wonders and terrors, we may find it difficult to imagine that profoundly simple wonder. Perhaps it is easier to imagine the wonder of the first people to hear a voice transmitted by radio – those attenuated, broken words out of the aether, riding a hissing, crackling wave. It sounds as if it comes from far away. Or hearing Bell’s first words over a telephone line, “Watson – come here – I want you.” Or encountering writing for the first time – “Words that stay,” as Jim Henson’s The Dark Crystal has it.
Most of us now learn words that stay when we are young, so young that the memory of the wonder of it has long faded. But watch a child understanding for the first time that a particular shape ‘makes’ a particular sound, and then learning to play with those building blocks.
Perhaps we can just about imagine the wonder of language itself – another set of signs, modulated vibrations in the air, shaped into words –“ox”, “pig”, “deer”, “rhinoceros”. You say the word, and at once, helpless to resist, I imagine the thing you have chosen to show me.
Before the painter of animals there was a painter of hand stencils; before that, a maker of geometric shapes, parallel lines, crosshatching. We will probably never know what they thought they were recording or communicating. But they felt that symmetry was important, and rhythm too. Something in those shapes was important enough to show to another.
Across millennia, we can encounter this art and experience some small echo of the wonder of contact with another mind. What they imagined was something like what we imagine. Not the same – their animal-headed, attenuated semi-human figures are strange to us, figures of nightmare now – but with enough in common to intrigue us, even across an abyss of time.
I wonder whether they told stories – I believe they must have, though I wonder whether, when words were new, stories can have existed as we know them, as deliberate fiction. I hope so. But in any case there was communication, and an attempt to express what one had experienced, and a hope, sometimes a desperate hope, that someone else would recognise that experience, and understand some part of it.
… And AI, of course, does none of this. There should be no debate really – at least, not yet – about the merit of AI visual art, or AI ‘storytelling’. The former is fun, at least for a time, though it becomes tiresome surprisingly quickly. The latter only has merit if you have no aesthetic sense of language: writing for people with no love for writing.
And the reason is absolutely straightforward: so far, while machine intelligence clearly exists, at least by some definitions, there is not yet a machine consciousness with a need to create art. No AI feels compelled to explore the profound mystery of itself, of the universe in which it finds itself, of other beings in that universe. It has no need to reach out to another mind and ask, Do you see what I see?
When human children realise they can make words, it soon becomes impossible to stop them – they have mouths, and they must speak. They want to tell you everything, and if they cannot make you understand, their frustration, sometimes their sadness, is profound. When they realise they can draw, they draw everything, and they want you to see, and they yearn for you to say that you recognise the image they are trying to put into your mind.
No AI experiences these, often overwhelming, emotions. Their art is for the most part only an echo of art they have been fed and told to imitate; their operators have perhaps imbibed the mantra ‘fake it till you make it’. But even faking it requires an idea of a mind with whom one is communicating.
None of this is to denigrate computer science, or the study of artificial intelligence, or even the generation of AI art. But if I thought AI art were the same type of thing as human art, I would have misunderstood what human art is, and what it is for.
It is telling, perhaps, that any discussion of AI art comes around very soon to the matter of profit. Most human artists, in every medium, would like to be paid for their work if they can – and why should they not? Many will continue creating art regardless of profit. But the key is that they would like to be paid because they have made, and communicated, something that has meaning to someone. Someone has encountered their art, and had a sense of understanding, or a sense of mystery, or simply an emotional response; they have been changed, even if only a little; with luck, their world is better for the encounter. And they may feel it fitting to reward the artist for that.
I wonder about those who seek to profit from AI art – what do they think art is, or is for? What has their own experience of art been? Perhaps they are moved by the beauty of the digital machine they are working with, as a steam engineer might be moved by the perfection of an engine. I hope so.
It is possible that a piece of AI generated art has moved someone; it is certainly likely that it has impressed someone. But if so, any meaning is solely in the mind of the observer. No communication of meaning has taken place. It is no different to my experience when I see a shape in the clouds – there, that is a bird, or a dog, or a person – my response is real, and may even make me smile – but it is my mind speaking to itself.
The sky is not lonely, and feels no need to speak to me.
And the steam engine feels no need to reach out to its maker.
There are no lonely machines.