FUTURE MANIFESTOS
KJØPMANNSGATA UNG KUNST, 9. APRIL – 31. MAI 2026
Åpning: 9. april 2026 kl.18.00
Kuratorer: Zane Cerpina, Espen Gangvik
Human Error [2025]
Human Error is a collection of oil paintings based on deliberately broken AI images centered on the human figure. Through a range of techniques, thousands of corrupted outputs were generated, and six were selected to be translated into paint. By painstakingly rendering these glitched images in oil and placing them on clean, white gallery walls, I recontextualize the “mistakes” as a deliberate aesthetic choice: an image artifact that is normally disposable, suddenly given weight, time, and permanence.
The bodies in the series hover between two emotional registers. On one hand, they are disturbing, grotesque in a body-horror sort of way – limbs repeat, joints misalign, skin seems to fold into itself, and the figure becomes a composite of competing drafts. On the other hand, they carry an unmistakable vulnerability. They melt, fuse, and soften rather than rupture. They are not presented as monsters, but as people caught in a kind of visual turbulence, mid-adaptation. That tension is central to the work: beauty and deformation are not opposites here – they are tightly coupled, the way they often are in synthetic imagery.
The distortions are not random. They are the result of intentionally pushing image models beyond their comfort zone so their internal logic shows itself. When you “break” an algorithm, it stops being invisible. It reveals what it privileges, what it struggles with, and what it tries to smooth over. In generative AI, the body is where this friction becomes most legible: hands, symmetry, touch, and the small cues that make a figure feel truly present. Human Error uses those failures as evidence – not to mock the technology, but to make its seams visible and emotionally charged.
All of the source images were generated using open-source AI models, mainly run locally. I worked in a way that deliberately invites the system to break down, selecting outputs where the model’s shortcuts and assumptions become visible rather than hidden. The aim was never to polish the images into seamless “AI art,” but to surface the artifacts – the moments where synthetic realism slips, and the underlying logic of the tool shows through.
French Girl, the largest painting in the collection, depicts a nude woman reclining in a blue bed. Her features are distorted by these AI glitches, but her gaze remains coherent. She looks directly at the viewer with an expression that reads as vulnerable and tired – an ambiguous mix of judgment and pleading, as if she is asking to be seen as more than a product of an image pipeline. The painting leans into that ambiguity rather than resolving it.
Her abdomen bulges unnaturally, accentuating an archetypal symbol of fertility and breeding. It can read as a kind of forced fertility, an exaggerated biological function, something closer to livestock than intimacy. At the same time, the rest of her face and body carry the “impossible” beauty standards that AI so efficiently distills – not because the machine invented them, but because it has been trained on decades of human taste, filtering, and repetition online. In that sense, her undistorted beauty is not neutral. It reflects what we have collectively taught these systems to optimize for, and the silent violence of that optimization.
Details in the painting underline that pressure. Her jewelry is composed of chained links. An armband sits on the body like a cuff. These elements are not a literal narrative, but visual cues: the figure is adorned and restrained at the same time. She is offered as an object of desire while also appearing trapped inside the very aesthetics that produced her. The work invites the viewer to stay in that discomfort. If the figure is judging us, it is not because she has moral authority – it is because she mirrors back our own image culture, condensed into a single, unstable body.
By turning a moment of algorithmic failure into a slow, crafted object, French Girl insists that the glitch is not just a technical error. It is a revealing event – a small collapse that makes the system visible, and a metaphor for humans trying to adapt to a world that accelerates faster than our bodies and identities can comfortably keep up with. In Human Error, losing coherence is not only frightening. It is also tender: a reminder that “being human” includes distortion, contradiction, and the need to be held together.
Syver Lauritz [NO]
Syver Lauritz (b. 1990) is a Norwegian multidisciplinary artist, designer, and creative technologist based in Oslo. Educated at The Oslo School of Architecture and Design (AHO), his work moves between physical craft and computational image-making, with a long-standing interest in what happens when technology bends, breaks, or reveals its seams.
After his master’s work (2015) explored “damaged” digital materiality through warped and broken 3D scans in a VR prototype, he went on to work across exhibition design, digital design, motion and animation, game design, and creative technology. He later served as a creative lead at TRY Dig, while continuing to develop a parallel studio practice.
Lauritz first became known to a wider audience through Portrettmesterskapet, reaching the semi-finals as a completely unknown hobby painter. Since then, he has held two solo exhibitions: his debut at Galleri HER (summer 2024), followed by Human Error at Grafills hus (April – May 2025). A third solo exhibition is scheduled in Oslo in May 2026.
Today, he runs an independent practice spanning commissioned work and self-initiated art, drawing on a background in design, marketing, and creative technology. He works across multiple mediums and formats depending on context – from commercial productions to studio-based painting.
