On the occasion of ArtVerona, the Carlon Collection, an eclectic and extensive collection of works from antiquity to the present day (including paintings, sculptures, engravings, drawings, miniatures, ancient books but also majolica, bronzes, ivories, furniture and decorative artefacts) started more than fifty years ago by the ìentrepreneur Luigi Carlon, presented its latest acquisition to the public: Borderland by Manuel Gardina (Brescia, 1990). The work, exhibited in a dedicated room at Casa Museo Palazzo Maffei, the prestigious home of the collection in the central Piazza delle Erbe, is a site-specific commission, born from the interest in science, innovation and new languages of a collector who has made the future the core of both his professional success and interests. It is impossible, based on this approach, not to engage with artificial intelligence and not to ask what are the prospects and contradictions that its increasingly broad availability can give to art. Borderland is a large sensitive digital canvas, which transforms through interaction with the public. Using motion and depth sensors to detect the presence of humans, the work, thanks to artificial intelligence, constantly reworks seventeen works from the collection selected by the artist, merging them into pairs, and then inserting its proposal for decoding the detected spectators into the resulting image in real time. A monitor supporting the environmental one in which the work lives shows the backstage of the generative algorithm, author of unrepeatable and constantly evolving scenarios, made even more immersive by a sound accompaniment that seems to translate the internal workings of the digital machine into harmonic polyphony. To delve deeper into the reasons for this pioneering work, we asked its author a few questions.
Manuel Gardina, “Borderland”, 2024
Emanuela Zanon: The reworking of the images of the collection that occurs in the work creates a singular coexistence of condensation and expansion. If on the one hand the algorithm tries to synthesize the salient features of the style of very different masters (such as Fontana, Magritte, Picasso, Canova, Balla or de Chirico) to return them through the lowest common denominator of a uniform “digital brushstroke”, on the other hand this restitution occurs in a dimension that projects the fictitious two-dimensionality of the canvas to infinity. What reflections or discoveries did this sort of coincidence of opposites inspire in you?
Manuel Gardina: Looking at the works in the collection, it emerges that the technology available to artists was often used to culminate in the creation of an object; with Borderland it was clear that I would have to focus more on the process. I remember the first experiments I did with generative artificial intelligence in 2018. It was an extremely crude tool, incapable of communicating with a certain degree of coherence with the operator, often producing questionable results. However, it was clear that what was stirring beneath those distorted and corrupted images was so distant from the reality of the time that it bordered on magic for me. I felt for a moment like the natives in Luis Sepúlveda’s “The world at the end of the world”, with the certainty of being in front of a tool capable of altering my immediate perception of the present. Speaking of perception, the paper “Attention is All You Need” (2017) comes to mind, which introduced the concept of Transformer, an architecture that has revolutionized natural language processing (NLP) and other areas of artificial intelligence. In parallel, human attention tends to focus on the most relevant stimuli, filtering out distractions. Similarly, the self-attention mechanism in Transformers “weighs” each part of the images in the collection and decides what to pay more attention to, ignoring irrelevant details. I wanted to create a work that gave us the ability to participate in this decision, so that our presence would have an impact on the way the AI processes and emphasizes the most significant elements of the collection. Everyone is inclined to give more importance to some elements than others, so I thought the concept of continuous evolution was central. The mechanism of re-elaboration of the works is constant, and so our attention to the work changes. I like to include this way of thinking in my process, it helps me escape the self-referential circle typical of my own medium. It also raises fascinating questions: how do we decide what is irrelevant? And, most importantly, who decides what deserves attention?
Manuel Gardina, “Borderland”, 2024
It is also very interesting the fact that AI, fed by a predefined set of data, tries to visually hypothesize what a human being is by comparing the audience that interacts with its sensors with its personal artistic memory. The result is disconcerting: in some cases the human features of figurative paintings, such as Boldini’s “La cantante argentina”, maintain an indisputable recognizability as such, while a real woman “scanned” by the algorithm can take on the appearance of a column or an even more abstract chromatic change of the existing landscape. What do these results tell us about the world in which AI understands us?
I share the point of view of the philosopher Luciano Floridi, according to whom the way of understanding the behavior of AI is that of a divorce: on the one hand, the undisputed ability to carry out complex tasks; on the other, the impossibility of attributing a meaning to its work. Despite the profound differences between us and AI, and the fact that we are still in the process of developing and understanding these technologies, it is remarkable to see that, for the first time in human history, it is possible to obtain such impressive results from a tool that essentially tries to simulate the behavior of our neurons. Being incorrectly identified and guessed by an AI, for example a column instead of a person, is often called a “hallucination” or a “dream”. I agree that it is useful to demystify the behavior of AIs as much as possible. What we observe can be compared to a cognitive bias, understood as distortions in data or models, which sometimes lead to unexpected and seemingly irrational decisions. This behavior throughout the history of media art can be observed as early as the 1960s in artists such as Vera Molnar and Roy Ascott, who explored similar dynamics. Their work demonstrates that a perfect process has never been necessary to create meaningful artworks. Similarly, with artificial intelligence and creative programming, chance partly replaces intuition, allowing the artist to devote much more attention to the process of programming the work (input) while letting the machine provide the results (output).
Manuel Gardina, “Borderland”, 2024
What kind of collaboration and hierarchy is established between the source works, the human being and the AI in Borderland and what does this triangulation predict about the way in which new technologies will influence our perception of reality in the future?
To imagine a possible future I would like to give a small example of how they influence us today. Let’s think about the warehouses of the American company The Feed, which occupy about 3,250 square meters and are designed to be completely automated. Dozens of robots, guided by artificial intelligence, collect up to 50,000 items ready for shipping every day. This trend is not limited to logistics, but extends to large agricultural supply chains and many other sectors. This shows that, instead of creating systems that act autonomously in the world, we shape the world so that they can function. We transform the environment so that it is compatible with computer systems. From this perspective, it becomes clear how technological evolution is progressively redefining the spaces and operating methods within our societies. If we take a step back and consider the context of the artworks, we can see how even a simple mixture of pigment can represent a technology capable of extending our capacity for personal expression over time. This is what makes the use of technology so engaging for me: despite everything, it always pushes us to think about what comes next, what is possible. Borderland, like much of the interactive and generative art, uses technology to fulfill a specific function. It is programmed to host depth sensors and an AI that analyzes and resizes the works in the collection multiple times pixel by pixel, as if to observe them from many points of view, until they are combined with a description. This allows the machine to focus on the aspects that most characterize the work. At the same time, being in front of the work, the spectator is never truly passive; he is a fundamental part of the creative process as his mere presence modifies the attention that the artificial intelligence places on the works in the collection. The spectator actively participates in the creation of the meaning of the work. This approach leverages the idea that enabling active dialogue with new technologies can be an extension of the ways in which we perceive our culture and how we construct new knowledge. It anticipates a future in which collaboration with artificial intelligence will not only be about efficiency or innovation, but will instead push us to reflect on how the mutual influence between humans and algorithms will affect the way we understand and experience reality.
Info:
Graduated in art history at DAMS in Bologna, city where she continued to live and work, she specialized in Siena with Enrico Crispolti. Curious and attentive to the becoming of the contemporary, she believes in the power of art to make life more interesting and she loves to explore its latest trends through dialogue with artists, curators and gallery owners. She considers writing a form of reasoning and analysis that reconstructs the connection between the artist’s creative path and the surrounding context.
NO COMMENT