I’m almost ready to sell my little booklet dialogues for one, and I wish I were starting this post by saying, yay, it’s ready!! It’s been delayed slightly, but all going well, I’ll have them in stock to start sending out by tomorrow or the next day. (You can pre-order in the meantime!) Despite that, there is a lot on my mind about AI and the questions the dialogue raises – not least of which is the (in my mind pointless and slightly bonkers) vitriol that emanates from photographers, of all people, towards machine-generated images. After all, what is a camera, if not a machine? But I’ll get to that in a bit.
I generated the pictures included in the booklet in the summer of 2022, soon after Midjourney was released. Some may recall a few blog posts in which I shared my earliest experiments. Those images were made using phrases from research for my MA, drawn from Freudian and Lacanian texts, some of which I understood, and many that were tricky to get my head around, to say the least. It wasn’t until I started reading Bracha Ettinger’s post-Lacanian feminist writing that I really stumbled upon something that felt ‘authentically’ me. “Authentic, using AI? Are you joking?” I hear some ask. Perhaps even Ettinger herself, since from what I have read, she’s not overly keen on digital materials – perhaps more of that another time. But, yes, authentic is the word I use. Not that authentic is everything it’s made out to be. Authenticity, like most signifiers, carries a multitude of inner and outer conflicts, questions, and shifting suppositions. These were explored by the Pictures Generation in the 1980s, and discussed in an excellent, highly quoted article by Douglas Crimp (1980). He writes;
“A group of young artists working with photography have addressed photography’s claims to originality, showing those claims for the fiction they are, showing photography to be always a representation, always-already-seen. Their images are purloined, confiscated, appropriated, stolen. In their work, the original cannot be located, is always deferred; even the self which might have generated an original is shown to be itself a copy.” (Ibid; p. 98)
I have used italics in the last sentence – because it seems to me that is the critical point.
In why is there an astronaut in a field of flowers/ I include the line, spoken by an AI companion, “The artist told me I’m a sort of photograph and I agreed but I don’t really know what she means.” Perhaps, if the AI and I were collaborating on the same stretch of writing today, instead of saying, I don’t really know what she means, the AI might have replied, “Aren’t we all, my dear?” Would such an arch response have been sophistry? Or is the question valid? And not only because we live in an era of reductionism, when nearly everything can and is routinely converted into fluid-like code, often so it can be placed on a graph and valued, even down to our aimless meanderings across the web, which are then turned into data, packaged up and sold? And is it still valid, while reality itself is being decoded, and can conceivably be recoded into any form? Even without any shiny technological advances, hasn’t Western culture long inferred that the vast majority of us are all a kind of image, a copy of some ideal original that exists elsewhere, out of reach; and especially far away and impossible to achieve if you’re not one of those accepted and feted rich white alphas?
Recently, a celebrated photographer submitted a machine-learning generated image to a photography competition and then, when he won the competition, he refused the prize with a dramatic turn, stating he had only done it to raise the point that AI images are not photographs. Well, he’s right, they’re not. But under the auspices which I suspect he operates, neither are most photographs nowadays. Even tin-types and other alternative photographic output might be rendered through Photoshop before being shared on social media, which automatically renders them as deconstructed code at some point. Personally, I don’t like the image that won. The photographer is right to refuse it. But for different reasons to the one he posits. There is something very telling about the state of photography competitions given it won, and perhaps it’s not the process that made the image which needs querying, but rather the notion that it is an interesting image in the first place.
So what of capturing time, one of the main tenets for which photography has always been known? After receiving proofs of the booklet, which include AI images made in 2022, images that are named <<Lamella (prompt-engineered) 2022>> (as can be seen here – scroll down to image no. 4 and here), I revisited Midjourney for the first time in months. I have been using RunwayML which provides me with a lot more flexibility and decision-making opportunity.
The difference was dramatic. It has become much more difficult in just a few months to subvert the apparatus, as encouraged by Vilém Flusser (2012) when he lauds experimental photographers because;
They [experimental photographers] are conscious that image, apparatus, program and information are the basic problems that they have to come to terms with. They are in fact consciously attempting to create unpredictable information, i.e. to release themselves from the camera, and to place within the image something that is not in its program. (loc 939)
As I have mentioned before, introducing anything to do with female production is fraught with trouble! It’s automatically understood as potential porn and rejected, resulting in blurry images via the Stable Diffusion plugin I have used in Photoshop. However, I seemed to get around this with Midjourney by using Ettinger’s complex and sophisticated psychotherapeutic language, much of which positions the womb as a central metaphor, as opposed to the phallus favoured by Freud and then Lacan. The images that emerged in 2022 are suggestive but manage to avoid being overt and actually pornographic. But now the same words elicit the same old unimaginative utilitarian lamella-like folds in nearly every image. Or if I use the name Hieronymus Bosch, as I did when generating those initial images, I end up with highly representative cartoon-like ideas of modern alternatives. Suffice to say, I feel like I captured a particular moment in the AI revolution which pertains to the moment of Midjourney’s birth onto the market. Even if I ask Midjourney to reference those earlier renditions, the outcome seems far less experimental and looks much like all the other AI stuff out in the ether.
This can be read in two ways. Photography is not the only recorder of time. But come to think if it, it never was. And subverting the machine has quickly become harder. Its tyranny is more tyrannical than ever.
Ultimately, I don’t think it matters if you use film, Photoshop or someone’s sperm to make an image. What matters is the work. To romanticise any media is to engage in a system of hierarchy, ideals and identity fetishisation. Which is why so much documentary photography which relies on old-fashioned snobbery to validate its existence, at the same time as claiming to query the status quo, only serves to reinforce it. And all those photographers who claim that AI is the devil incarnate, might like to read what people were saying about photography back in the day when it first emerged!
Do buy my little booklet if you can. While writing this, I heard it should arrive tomorrow. And the proceeds will help pay for my degree show!
Images made in 2022 using Midjourney, some of which appear in dialogues for one
Images made with the same text referencing the older images created in 2023 with far too much form which is hard to escape, and a page from Capturing the Light by Roger Watsdon and Helen Rappaport (2014)
Midjourney’s offerings using the same prompt to make the images I’ve included in dialogues for one
Crimp, D. (1980) ‘The Photographic Activity of Postmodernism’ In: October 15 p.91.
Flusser, V. (2012) Towards a philosophy of photography. London: Reaktion Books.
Rtology, E. (2023) AI Generated Art Rocks the Photography World: Award Winning Image Sparks Debate. At: https://medium.com/data-driven-fiction/ai-generated-art-rocks-the-photography-world-award-winning-image-sparks-debate-ca567bce99c (Accessed 17/04/2023).
Spelling/grammer corrections made on 21/04/23