"If there is something that opens the horizons, it is precisely ignorance."
I read this sentence by writer and diplomat Romain Gary during a train journey from New York to Washington. It takes me a few minutes to interpret the sentence taken out of its context.
I wonder if the term “ignorance” in the phrase should be interpreted as "non-knowledge" and, as such, if the status of “not knowing” could potentially stimulate curiosity and thus open the horizons, metaphorically speaking, of the mind.
I reverse the order of factors and imagine if horizons can then fight ignorance.
It is a tortuous thought, the one going on in my mind. I wonder if it is happening because, while reading, from the window on my left, the flat-blue and moderately misty horizon of the Gunpower river bay opens to my view or because I lack the right amount of morning caffeine.
The landscape in front of me is not extremely exciting. Still, it reminds me of a photo taken a few years earlier from the shore of the river Ob in Russia's Yamal Peninsula during a freezing -42c morning in 2019. The river, in its vastness, was a pale blue frozen horizon crossed by a dark blue glare.
The image is part of a work about gas extraction in the Yamal Peninsula. Over time it has also become one of my best-selling edition prints, although most buyers are, I believe, unaware of the motivations and context in which the image was made.
By a tortuous generative association of thoughts, I connect Gary's sentence to this image - aided by the window landscape – and to an article by Lev Manovich on generative algorithms that turn textual input into images that I have read a few weeks before.
While waiting to arrive in Washington, I uploaded “River Ob” into the DALLE-2 software and tried to describe the photograph.
In"Automating Aesthetics: Artificial Intelligence and Image Culture", Manovich explains that: "While today AI is already automating aesthetic choices, recommending what we should watch, listen to, or wear, and also playing a big role in some areas of aesthetic production such as consumer photography (many features of contemporary mobile phone cameras use AI), in the future it will play a larger part in professional cultural production.”
The first attempt of producing an image starting from text description is simplistic. I purposely start with a single word to test the system, and I am not surprised that the essential input 'river' gives stereotypical aesthetics and quite disappointing results.
I try to describe my River Ob image and enter: "frozen river in Russia". Again, the image generated has no similarities with the one I took and have in mind. It is, again, in my view, sad and insignificant.
With some additional effort, one term outlines the resulting turning point. The key word is "horizon", and the new image is now incredibly similar to my original photo.
I repeat the process with minor, aesthetic descriptive variations by adding my original photo as a parameter to the last "satisfactory" result. I run the command to generate 100 versions of River Ob, in which each iteration takes the previously generated image as a parameter.
I obtain a sequence of variations in colours, shades and slightly different compositions. They are vaguely reminiscent of Anni and Josef Albers' colour studies with a mood of Rocko's canvases. I’m a big fan of both.
I start making spontaneous connections between the theory of technical images and the resulting images. In the meantime, my thoughts divert, trying to make sense and links between those first thoughts and the sentence I read in the book, the pictures generated by the algorithm, my emotional state when I received the book as a present, the landscape that passes before my eyes, the trip to the Yamal peninsula…
What could Walter Benjamin's reaction be facing this technological progress applied to images? Suppose the aura - 'hic et nunc' (uniqueness and unrepeatability) of the work of art was already compromised at the time of its mere reproducibility. How would he have judged this new 'generative' capacity? And what would Charles Peirce and Paul Levinson have thought about the indexicality of photography today?
It seems evident that when photography begins to move away from the etymological definition of "writing with light", it also emancipates itself from the restrictions of the context that has produced it. It also outdistances itself from the subject, the time and the space it represents. It becomes, perhaps, pure creation, whoever or whatever is the performer of this creation.
I will not try, at least for now, to go any further on the philosophical theory of image creation. Yet, Roland Barthes' now famous photographic paradox: - "what the photograph reproduces to infinity has occurred only once: the photograph mechanically repeats what could never be repeated existentially" - seems to define quite well the mechanical-generative exercise produced by my primordial input and the few seconds of work of artificial intelligence.
Still, the 100 River Ob are, in this case, not only copies but 100 interactions and variations, in some way connected to an "existential" moment.
And here I begin to spiral myself on that word: 'existentially'.
Existentially, these 100Ob are the result, albeit mechanical, of my first, real, lived, interpreted experience of 'being' physically there in that Russian winter at -42. They are also the product of a verbal and textual description that is not simply content-based but aesthetic and, therefore, emotional, at least in a broad sense.
Why did I take that photo, choose it, when I edited it, and now for this experiment while reading this book? Why now? Why, above all, I chose these words to describe it:
"A river
At dawn, in the cold of a pale blue
Ahead of me, a distant horizon
A line at the infinite centre of my gaze.”
I do not, at least for now, have a detailed answer to any of these questions. Still, as I look at the images produced, I try to perceive an emotional value in each of them, and nothing brings me back to the feelings of the 'original' image.
Yet there is something magical that aligns my mental connections between the image, the book, the journey, the memories and the attempt to find an explanation for these connections that inevitably associates with the mathematical process performed by the 'intelligent' machine.
The elementary definition of Artificial Intelligence is "computers being able to
perform many human-like cognitive tasks. If applied to the creative, artistic or aesthetic process, it could be defined as: "humans programming computers to create with a
significant degree of autonomy new artefacts or experiences".
Suppose it is true that the emotional response is the key to exposure to artistic experience. In that case, it is undoubtedly subjective that my reaction to the result of 100ob has little to do with the mechanical process performed by the algorithm but rather with the mnemonic and receptive connections that my brain triggered to produce them and the considerations that the result generated.
At the same time, both the trigger and the result have no value except for the process of the Pindaric flight that occurred in my brain.
The literature on Ai often attributes to it the social, ethical and moral imperfection of being a 'black box' from which it is difficult to extract the criteria and details that produce a given result. Yet on a commercial and personal level, the uses are the most diverse and increasingly in contexts that are more and more associated with actions- reactions linked with human emotions.
Students write essays on it to pass exams. Companies use it to understand tastes, sell products, and produce eye-catching copy and images. Some use it to write resumes or biographies, poems, and one friend - he confessed to me - successfully used it to reply to chat messages on Bumble.
In almost twenty years of 'artistic' practice, albeit documentary, I have, with careful conscience, always tried to distance myself from the desire to express the slightest form of personal emotion in my images. I have always supported the Dusseldorf artistic approach to preserve a record of neutrality of the images and to hide my emotional vulnerability from external judgements. I have, in fact, carefully guarded my creative and mental processes. I concealed it in aesthetically cold results and hid it within a 'black box' not dissimilar to automated intelligence.
I wonder if, in part, and at different levels of mimicry, this is how all humans’ brain works.
How often is it challenging to interpret human nature, the thoughts of our loved ones or those we work with? How often does it seem impossible to understand the thoughts of others or even to listen to and understand the sequence of those thoughts that we produce on our own and develop reactions and results?
You only have to read the famous book “The Man Who Mistook His Wife for a Hat “ by Oliver Saks to understand how the complexity of the human mind is far more abstruse than artificial intelligence, albeit governed by incomprehensible but precise chemical and mechanical rules and, sometimes, mistakes.
This frivolous experiment on a train, sparked by a sentence in a book, the sight of a landscape, the memory of a place, the simplicity of a technological tool, and all the thoughts that followed, are not valuable for the resulting images that were produced. Instead, it was intriguing for me and only for me, for the cerebral process that trigged.
I wonder if automated generative processes can help produce emotional and rational thoughts or trigger ignorance to open horizons.