Oggi ho deciso di rinfrescare e riadattare un pezzo complesso, un pezzo tormentato, faticoso, vittima indifesa di quelle vicissitudini che accomunano le cose vive. Dovrei cercare di non dilungarmi eccessivamente in una brodosa introduzione, ma è giusto contestualizzare: il testo qui sotto nasce da un format che avevo ideato originariamente per un blog di settore (Design Disciplin) con il quale un caro amico mi aveva messo in contatto qualche mese fa. Per diverse questioni nel merito delle quali non entrerò, ho sentito di voler tenere queste interviste con me, nella mia umile celletta dei server di Substack.
Il format è semplice e consiste in un'intervista non strutturata ad artisti, grafici e figure simili rispetto all’utilizzo delle intelligenze artificiali nei loro settori. In primo luogo ci tengo a scusarmi verso tutti i personaggi coinvolti che si sono dimostrati entusiasti e disponibili rispetto alla mia proposta, dal momento che inizialmente avevo comunicato loro che sarebbe uscito su DD, quando invece ho gelosamente costudito le loro sagge riflessioni. Successivamente, ci tengo invece a ringraziare Fabio Cassisa, il celebre “caro amico” che oltre a offrirmi la possibilità poi rifiutata, mi ha dato una mano nel brainstorming e nella rifinitura dell’articolo. Infine devo porgere le mie scuse ai lettori esclusivamente italiani, perché l’intervista è stata tradotta interamente in inglese fin dalla prima stesura.
Il primo episodio è il flusso di Giovanni Lo Castro, brillante artista romano che ho avuto il piacere di conoscere qui a Milano, dove attualmente studia e lavora. In coda all’intervista troverete un profilo più dettagliato di Giovanni, che ancora ringrazio veramente tanto per il content di altissima qualità che ha generato con la nostra chiacchierata. Detto questo, non mi resta che augurare a tutti buon divertimento e un abbraccio a voi che, fedelissimi, mi leggete ancora <3 <3
“Technology has really impacted my art and I would say that it has played a major role in developing my style. My first approach to art was through the study of painting: even though this medium is still core to my artistic process, digital tools such as photoshop made me realize how my interests were far beyond that. Today, I can manipulate images in many groundbreaking ways, which has been made possible by technology, yet the biggest revolution wasn’t even this one: the image manipulation process through digital software led me to understand the physical aspects behind analogical images as well. Facing human ways to summarize notions in a reference system drew my attention to the broader field of perceptive aspects of knowledge: that’s when I started looking at images not merely from a frontal, sometimes trivial, point of view and I went on to meta-analyze them. What am I seeing? What does seeing mean? How can I make the viewer appreciate this point of view? Technology has been very central to the undertaking of this path.
We sense images primarily in a digital way, so, when I am watching images through a screen, what am I actually observing? That’s how I learned about Bayer grid and RGB, the analyzable, complex system behind screen images: this allowed me to specify some closer dynamics of the image viewing, such as focusing on the image content or the container instead. Digital images are nothing more than a pile of bits read by a software. Just data. We are always observing the shade of the image, a subsequent reconstruction made out of microscopic red, green and blue dots. Adding AI to this process means adding a huge increment of invisible data: technology, on one hand, discloses something out of the hidden, while, on the other hand, conceals a great part of his work.
The majority of image generators AIs are text-to-image, therefore natural language is key such as a brush to a painter. This communication way we are using with AI speaks to the hidden archived of humans, the collective unconscious treasured in the internet through data. That is why I think that approaching generative AIs is a self-reflexive exercise for society. This is clear when we think of databases such as ImageNet or FFHQ: they all reflect the biases of the people and they return the average of it. It’s similar to telling your friend about a dream: AI tries to reassemble its creation with confusing, sometimes even invented, elements and there are emotional responses that come out of the recognition of something close to your oniric idea of the representation.
I think that AI will end up being just a research tool despite today's hype: I don’t want to consider the product of AI definitive. The result of a prompt cannot be worth as an art object, it would be like considering satisfying the first google result to write a whole thesis. We should become more aware of this fact: an AI generated image is an unrefined object which should be wrought by the artist. Nowadays, it's pretty easy to spot AI-authored content. If we don't stay away from it, we'll end up with mass-produced, inflated art. As I pointed out earlier, AI will become just one of the many sources an artist could use for inspiration.
However, if the AI-generated image is the result of a prompt born out of an investigation, then there are exceptions. My next AI-based art project is trying to work on remembrance: I would like to involve friends who share some memories with me in describing it to the AI in order to obtain the missing picture of that moment. Here, the artistic process is not delegated to the machine but it shifts before the generation. I already tried once with my cousin and we reached a surprising result. Our description of it was so detailed that it came out as an absurdly accurate portrayal of the location, event, and emotions we were involved in. Even the aesthetic was reminiscent of an old phone-picture. This recall aspect of AI is great and I will continue going down this path, interviewing people to obtain a shared picture out of memories.
The long-standing question on what can be considered art nowadays is more discussed than ever. Digitalization, blockchain contracts and NFTs made it very hard to distinguish what could be considered art or not. Is a million dollars sale enough to give a piece the art status? Or is art only what aesthetic concerns? I think that approaching this problem from an economic perspective can easily be misleading. Personally, I consider AI-generated art worth the status whenever the piece comes from artists who are working on their relationship with the image. There’s the need for awareness, just like abstractionism: everyone could paint something abstract, but art kicks in when you’re consciously telling something about the world or about yourself with that work.
My transition from painting to digital art has been already driven by frustration. I asked myself what was the point of creating new images when the number of them on the internet is steeply high and quickly rising too. Today’s artists have a frustrating relationship with images regardless of AI: art needs to go over the medium and carry an artistic, derived from feelings, code instead. The focus necessarily shifts from the content of an illustration to the process and manners of that piece: even if that image can still be reproduced and generated ''as if'', it lacks tangibility.
AI is not that far from the human way of thinking. Actually, I think it can be reduced to a human brain appendix: this idea gets stronger when you realized AI is pointless without someone taking advantage of it. We use creative AI specifically to boost imagination, but it’s not the only cognitive function AI can enhance. As a matter of fact, AI won’t produce anything without injecting human data in them: this makes them comparable to our cerebral way of creating thoughts. Though, a big difference shows when we enjoy something: a machine is not capable of appreciating the actual image content as we do. AI can merely look at images from the metadata point of view.
Lastly, this anthropocentric reading of technology led me to question how this can be ecologically sustainable. The energy required to power this system up is constantly growing: we need actively functioning support to generate and archive everything existing in the digital world. We need real, physical matter to make this possible. A hard disk is nothing but a new kind of book: it has stiff limits, both in space and time. This doesn’t concern us in our everyday life compared to the other forms of pollution, yet there’s someone worried about the topic. We should keep an eye on this while we talk about everything that’s tech-related.”