The Wrong Biennale is a decentralized digital art biennial built through online pavilions and offline embassies rather than a single fixed venue, and Aisthesis Lab joins that structure with a curatorial program centered on digital materialities, hybrid architectures, mixed realities, cyberculture, and the convergence of space, image, and information. Metatopia sits at the meeting point of those two frameworks. It belongs to a biennial shaped by networks, screens, and distributed viewing, and it is hosted by a platform already invested in how technological media alter perception, embodiment, and spatial experience. The current edition, “Delirium Ex Machina,” is framed through AI hallucinations and a break from deterministic machine logic, which makes the exhibition especially suited to a reading focused on unstable perception. What emerges from that context is not a celebration of machine intelligence as a more advanced form of vision, but a sharper examination of what happens when technological systems try to classify, decode, and expose the world with too much certainty.
The Wrong Biennale has built its identity around a form of exhibition making that belongs to the internet age, not only because it presents digital art, but because it understands circulation, fragmentation, distance, and screens as part of the condition of contemporary viewing. Rather than gathering everything into a single monumental venue, it unfolds across a distributed structure of online pavilions and connected spaces, allowing artworks to exist within the same networked environment that shapes so much of their language. Aisthesis Lab enters that structure with its own distinct focus on digital materialities, hybrid architectures, mixed realities, and the shifting relation between image, space, and information. That combination gives Metatopia a very specific force. The exhibition emerges from a biennial built for networked art and from a platform already committed to thinking about how technological media reshape perception. Framed through “Delirium Ex Machina,” the project turns away from neat fantasies of precision and instead moves toward instability, hallucination, and the strange point where systems built to decode the world begin to expose their own uncertainty. From the beginning, the exhibition does not treat machine seeing as a polished achievement. It treats it as a restless field where the effort to classify reality starts to slide into distortion.
That idea comes into especially sharp focus in Jiaqi Liu’s Only Cloud Knows, a work that cuts directly into the authority of machine recognition. Built as a Chrome extension using Google Cloud Vision AI to perform art recognition, the piece overlays confidence labels onto images, making visible the system’s attempt to transform visual ambiguity into measurable certainty. What is so striking about this work is that it does not need a spectacular glitch to show instability. The instability is already present in the confidence itself. The image is approached as something to be named, ranked, and translated into legible output, as if the visible world were waiting to be reduced into proper categories. The work reveals how much violence can hide inside that apparently neutral process. A label does not simply describe an image. It narrows it. It turns perception into a form of capture. It assumes that once something has been properly tagged, it has been understood. By foregrounding this logic, the piece becomes more than a critique of one recognition system. It opens up the deeper problem running through the exhibition, which is the desire of technological vision to make the world fully readable and the way that very desire produces a damaged form of sight.
Apolinário’s an insect robot preparing a meal series pushes the same problem into a more absurd and synthetic register, which makes it all the more revealing. The works are organized around prompts that correspond directly to their titles, using language as the starting mechanism for generated imagery. On the surface, this sounds almost straightforward. A phrase is given, and the system produces an image. Yet what appears is never just the clean delivery of a command. Instead, the resulting scenes feel exaggerated, unstable, and slightly deranged in their certainty. The prompt asks for something specific, but the image seems to overread the request, filling it with ornamental excess, stylistic contamination, and a theatrical intensity that makes the whole scene feel both obedient and out of control. That is what makes these works so useful for understanding the exhibition. They show that synthetic image systems do not simply mirror language. They translate it through a process of visual inflation. The result is not clarity, but a kind of delirious exactness that reveals how technological seeing often operates by taking the world too literally and too strangely at once, producing images that expose the gap between command and perception rather than closing it.
Vadim Epstein’s works deepen this atmosphere of unstable recognition by keeping close to the human image without ever fully securing it. Pieces such as Kitezh, They, Subcultural genetics, and Virgo Latenta emerge from a practice centered on generative methods, figurative imagery, machine learning, and GAN-based video. What is so effective in these works is their refusal to let the figure settle. Human likeness remains present, but it never reaches full stability. Faces appear to form and dissolve at the same time. Bodies seem to announce themselves while slipping away from fixed identity. The viewer is brought close to recognition, but not allowed the comfort of certainty. This is where the exhibition’s theme becomes especially rich, because the problem is no longer one of obvious malfunction. It is one of continuous approximation. The machine does not fail because it cannot make an image. It fails because it keeps making images that hover on the threshold of the known without ever grounding themselves fully. These are not blank outputs or broken files. They are persuasive and unstable presences, synthetic figures that show how technological seeing can generate resemblance while remaining estranged from the very reality it claims to model.
Denis Volnov’s datamoshing works shift the argument from generated figuration to corrupted movement, showing that unstable perception is not confined to still images or AI hallucinations. His practice manipulates appropriated pop-cultural material through glitch processes that exaggerate gesture, body language, and motion until the image begins to leak, smear, and convulse. In these works, continuity becomes visibly fragile. A figure does not move from one point to another in a clean sequence. It drags pieces of itself through time. It leaves residues. It becomes a body reconstructed through compression errors and broken predictions. That transformation matters because it reveals what smooth digital vision usually tries to hide. Every moving image built through technological systems depends on segmentation, substitution, and fragile temporal stitching. Datamoshing drags those hidden mechanics into the foreground and turns them into an expressive language. The result is often funny or grotesque, but never merely decorative. These works show that technological images are always held together by unstable procedures, and that the promise of seamless machine perception depends on suppressing the very operations that make it possible. Once those operations surface, the authority of the image begins to fray.
Nikita Diakur’s simulated works make this instability even more concrete by focusing on how machines attempt to model bodily action itself. In Fighting Gravity, a machine learning model learns to make a model human stand up, turning one of the simplest human gestures into a prolonged episode of awkwardness, correction, and repeated failure. The humor of the piece arrives quickly, but it is not shallow humor. What it exposes is the distance between embodied life and computational procedure. Standing is ordinary for a person because the body carries habits, balances, and sensorimotor intelligence that do not need to be consciously solved each time. For a system trying to simulate that action through rules and iterative learning, the same gesture becomes unstable, clumsy, and visibly difficult. This is where the exhibition becomes especially sharp in its treatment of technological vision. Seeing is not separate from modeling. To render the body, the system must also make assumptions about what a body is, how it behaves, and what counts as successful motion. Diakur’s work reveals that behind the polished surface of digital simulation lies a fragile process of approximation, one that exposes how poorly machinic systems can inhabit the physical realities they attempt to reproduce.
John Bumstead’s Broken Screen of the Day series grounds these questions in hardware and material support, making it impossible to imagine technological sight as something immaterial or purely abstract. The works are made by photographing images through broken LCD screens and then rephotographing the results, allowing cracked crystals, damaged displays, and compromised signal paths to reshape what becomes visible. This move is crucial within the exhibition because it insists that the machine-made image is never just an algorithmic event. It is also a physical event, one that depends on surfaces, chips, displays, wiring, and all the vulnerable infrastructure through which an image appears. The damage in these works is not a superficial effect added for texture. It is evidence that the image is always passing through unstable matter. That matters for the larger argument because it shows that the uncertainty of technological vision is not only ideological or computational. It is also physical. The screen can break. The signal can fail. The display can contaminate what it shows. These works bring the exhibition back to the objecthood of digital media and make clear that machine seeing is always shaped by the limits and fragilities of the devices that carry it.
The exhibition becomes even stronger when these tensions are placed alongside works that move through hybridity rather than overt breakdown. QNO’s sculptural pieces, including Peugeot 106 twisted and works from the Solid Crystal Display series, translate digital form into resin, wool, paint, and polystyrene, refusing the separation between synthetic aesthetics and tactile matter. Dekso’s the Walker imagines the real and the virtual as two sides of one visible field, giving the human figure and its surroundings a hybrid visual status that never resolves into a clean division between physical presence and digital construction. Read in relation to the more obviously generative, glitched, or simulated works, these pieces show that instability does not always appear as collapse. Sometimes it appears as unresolved fusion. The image does not shatter, but it cannot fully decide what kind of object it is. The figure does not disappear, but it no longer belongs securely to one order of reality. These works widen the exhibition’s scope by showing that the machine-made image is not only a site of error. It is also a site where categories that once felt stable, such as virtual and material, generated and embodied, become increasingly difficult to separate.
What finally gives the exhibition its shape is the consistency with which it returns to the impossible ambition behind contemporary technological sight. Across these works, visual systems are tasked with recognizing, simulating, classifying, translating, and exposing the world, yet again and again what they produce is drift rather than mastery. Labels overreach. Prompts mutate into excess. Figures approximate without settling. Motion leaks through compression. Bodies resist simulation. Screens mark the image with their own damage. Hybrid objects refuse to stay in one ontological category. None of this feels accidental. The instability running through the exhibition is not just the side effect of imperfect tools. It is a revelation of the deeper logic beneath them. These systems are built on the fantasy that reality can be made fully legible through technical means, and the works gathered here repeatedly show that such a fantasy generates its own distortions. The more forcefully machine vision tries to capture the world, the more visible its own strain becomes. What Metatopia offers, then, is not a simple warning against technology or a celebration of digital weirdness for its own sake. It offers a sharp image of a present in which technological power remains immense, but the certainty once promised by its gaze is already coming apart.
Visit https://www.aisthesislab.art/ until March 31st, 2026. Metatopia includes work by Alex Manea, Alexandra Bouge, Alexander Limarev, Amal Alshoura, Andreas Koens, Andrew Reach, Anne Herzbluth, Aphex Redditor, Arad Kljshu, Arnau Tàsies, Bariya, Beatrice Lartigue, Beile, Blanche the vidiot, Bob Georgeson, BSBLOrk, Carla Lombardo, Christie Lau, Clear Shadow, Das Vegas, David Mew, Davis Lisboa, Digital Martins, Elena Romenkova, Elif Sezen, Emanuele Dainotti, Emma Cosgrove, Enco, Erik López, Eris Spam_, e-topia, Fabiola Larios, Flávia Goa, flvz_, Fu Wenjun, Fúcsia, Gayatri, Gabriel Pessoto, Gizella Popescu, Gopakumar, Group 4, Guaraci Nanferdes, GUOFEI, Ian Benjamin Callender, Ivana Tkalčić, Jes Chen, Jiaqi Lu, Jonas Esteves, Josephine Florens, Julia Rocha, Karen Eliot, Katya Kan, Ksenia Kudasova, Lara Peters, Lev Manovich, LfV1ÄT4, Leonardo Matsuhei, Longdan Yan, Lerabo, @loveletter.exe, Maciek Stępniowski, Madam Memoticon, Malitzin & Ivan Abreu, Marta Di Francesco, Matheus Solar, Mateo Campulla, Mehreen Hashmi, Michael Woodruff, Moksha Kumar, Mônica Amêndola, Ms.V, Nicolas Tilly, Nikolina Schuh Netz, Nina Sobell, Nirali Lal, Ogg Lullenstyd, Ole Tersløse Jensen, Philip Wood, Pomba Molex, Rhett Tsai, Ricardo Nolasco, Rita Raeva, S4RA, Sabrina Menedotti, Sailor Noom, Sandrine Deumier, Syporca Whandal, Stephen Roddy, Sue Nhamandu, Tassia Mila, Tasha Lizak, Tati Cocteau, Tripura, TU Lang, Tyler Kline, Udi Cassirer, Valdas_NeuroVirtual, VFCC, Vicent Tanguy, Vittorio Bonapace, Vladimir Nefedov, Xátana Potyguara, Waterflower, Winnie Soon, Yasaman Sharifzadeh, Yi Tang, Yves Gregoire Lizárraga, Zander Porter, Zkymicx. It was curated by Luciana de Paula Santos from Aisthesis Lab in São Paulo, Brazil.
