Between Glitch and Gesture
dawn jian’s paintings are most compelling not when they visualize artificial intelligence, but when they force digital imagery to surrender its apparent frictionlessness. Generated fragments pass repeatedly through pigment, canvas, scraping, gravity, and bodily gesture, accumulating errors and resistance along the way. In this sense, jian’s hybrid process is less a seamless fusion of human and machine than a productive failure of translation. Chinese ink sensibilities, Western abstraction, and computational imagery remain visibly unresolved, allowing the paintings to register the contemporary sublime not as technological spectacle, but as a condition of attraction, instability, and material consequence.
From ArtVanguard Editorial Team
Artist Biography
dawn jian is a Los Angeles–based multidisciplinary artist working across painting, generative AI, digital image-making, and material fabrication. Born in Guangzhou and raised in Canada and the United States, they draw on a cross-cultural visual language that brings Chinese ink sensibilities into conversation with Western abstraction and emerging technologies.
dawn creates custom-trained AI models using their own paintings, translating generated fragments back into physical works through gestural painting, collage, scraping, and layered surface treatments. Their practice examines how technology reshapes perception, authorship, and the natural world. Glitches, pixelation, eroded surfaces, and mineral-like textures connect the instability of digital images with the material and ecological infrastructures that support them.
They received their BFA and MA from the Maryland Institute College of Art. Their work has been presented in solo exhibitions in Hong Kong, as well as in group exhibitions and art fairs in Beijing, Shanghai, Hong Kong, New York, and San Francisco.
Artist Website: https://www.dawnjian.com
Instagram: @okidawwwn
Artist Statement
My practice moves between painting, digital image-making, and installation, examining how images and cultural memory are transformed as they pass between physical and computational systems. I train custom LoRA models on an archive of my own paintings, then use CivitAI, ComfyUI, and Photoshop to generate, distort, and recombine visual fragments. These fragments become provisional compositions that I translate onto canvas through staining, impasto, layering, and scraping.
Scraping is central to this body of work. As a physical gesture, it removes material, exposes buried layers, and leaves scars across the painted surface. Conceptually, it echoes the way artificial intelligence systems scrape images, text, and cultural memory from the internet and convert them into training data. The canvas becomes a record of both material erosion and algorithmic extraction.
Pixelation, compression artifacts, and glitches interrupt the paintings like corrupted signals. These visual disturbances evoke a world in which landscapes erode alongside our diminishing attention spans, while images are repeatedly compressed, reformatted, and circulated through screens and digital platforms. They point to broader transformations in perception, increasingly shaped by machine vision and the technologies through which we encounter the world.
Through these hybrid surfaces, I approach artificial intelligence not only as a method of image production, but also as a system entangled with ecological extraction, political power, and economic infrastructure. The work asks what happens to perception, memory, and material experience when images are increasingly produced and formatted by machines.
Interview with dawn jian
Q1. You were born in Guangzhou and raised in Canada and the United States, and your practice now moves across different international art contexts. How has moving between cultural contexts shaped your relationship to visual memory, language, and image-making?
I think my hybrid upbringing has shaped how I understand images and the world around me. Moving between cultures has made me more attentive to broader, more universal ideas and experiences, while also giving me the freedom to combine details from different visual languages and processes. After several years living and working in China, I became especially conscious of the speed and pervasiveness of digital culture and the growing presence of AI, which I wanted to infuse into my paintings upon returning to the United States. At the same time, living in major cities made me sensitive to density, grit, noise, and visual overstimulation, often leading me to seek their opposites: stillness, openness, and the spaciousness of natural landscapes and horizons. My paintings often try to hold these tensions in balance, moving between emptiness and accumulation, stillness and acceleration, and natural and technologically mediated ways of seeing.
Moving between international art contexts has made me attuned to how meaning shifts across cultural assumptions, values, and ways of seeing. A word, gesture, or image can carry very different implications depending on its context, and I’m interested in what gets lost, altered, or misunderstood in that process. That sensitivity to translation has shaped my interest in ambiguity and in images that can hold multiple readings without resolving into a single meaning.
Q2. Your work brings Chinese ink sensibilities into conversation with Western abstraction and emerging technologies. How do these different visual languages meet in your practice, and what kinds of tensions or possibilities emerge when they are brought together?
My practice looks for ways to bring traditional Chinese brush painting into conversation with Western abstraction through shared interests in gesture, chance, materiality, and ambiguity. I draw on techniques such as 没骨画法 (“boneless” painting), 泼彩 (splash-color), and 留白 (activated emptiness), where washes, flowing pigment, and unpainted space can create atmospheric depth and a sense of incompleteness. Alongside these traditions, I work with embodied gestures, soak-staining, smearing, scraping, and unconventional tools such as mops, sponges, and squeegees, drawing loosely from the histories of American Abstract Expressionism and action painting.
Rather than trying to merge these traditions into a seamless language, I am interested in seeing what happens when their differences remain visible. Chinese brush painting often values restraint and suggestion, while Western abstraction has frequently emphasized physicality and accumulation. In my work, I try to navigate between delicacy and aggression, control and surrender, while making room for controlled accidents. I hope the paintings can become a space where different approaches to image-making coexist without needing to be fully resolved.
dawn jian, UnPressed, 2026. Acrylic, fluid acrylics, metallic acrylic, and dye over custom AI-generated digital image collage on archival pigment print on canvas, 36 × 24 in.
dawn jian, TouchFlow, 2026. Acrylic and fluid acrylics over custom AI-generated digital image collage on archival pigment print on canvas, 36 × 24 in.
Q3. Your background includes Environmental Design, Social Design, and Painting. How has this path shaped the way you approach a painting as something spatial, constructed, and process-based?
My background in Environmental Design taught me to pay close attention to the physical conditions of a site: scale, light, architecture, and the way people move through a space. Social Design made me more aware of intangible conditions as well, especially cultural context and the different ways people might experience the same environment.
I try to bring both sensibilities into my painting practice. When possible, I look at the exhibition space early on so I can imagine how the paintings might live there, rather than thinking of them only as isolated objects. Since I often curate my own exhibitions, I also consider sequencing, spacing, and how the works relate to one another. I try to remain aware of the questions circulating through contemporary art and culture and allow them to inform the conceptual framework of the work.
Q4. You describe your practice as examining how “the experience of the sublime” is changing in an age increasingly shaped by artificial intelligence. What does the sublime mean to you now?
The sublime has traditionally been described as an encounter with something so vast or powerful that it exceeds our ability to fully comprehend or control it—often in relation to nature. Increasingly, I find that same mixture of awe, anxiety, and disorientation in our relationship with technology. AI and digital infrastructure have become so embedded in everyday life that they can feel almost like an ecosystem: enormous, interconnected, and largely invisible.
My paintings explore this idea of a digital sublime, but I’m interested in keeping it grounded in material experience. I train custom LoRA models on an archive of my own paintings, then generate, distort, and recombine fragments before translating them back onto canvas through staining, impasto, layering, and scraping. Scraping has become particularly important to me because it operates in two ways: it refers to how AI systems scrape images, writing, and cultural memory from the internet as training data, while also existing as a physical gesture that removes material, exposes hidden layers, and leaves scars on the canvas.
For me, the sublime now has to do with the scale and speed of technological change. Technology is reshaping both the physical world and the way we perceive it, and I use pixelation and glitches as forms of visual interference. They suggest a world in which landscapes are eroded as quickly as our attention spans. Images are also constantly degraded, compressed, and reformatted by screens and machine-vision systems. In that sense, the work is not only about AI as an image-making force; it is also about the larger infrastructures that shape this new visual reality—and are shaped by it in return.
dawn jian, Recoded I & II, 2025. Acrylic and custom AI-generated digital image over archival pigment print on canvas, 35 × 70 in.
dawn jian, Brainrot Me Not, 2025. Acrylic and custom AI-generated digital image over archival pigment print on canvas, 39 × 28 in.
Q5. Your practice often begins with custom-trained AI models based on your own paintings. What first drew you to training AI on your own visual archive, and how does an AI-generated fragment eventually become a physical painting in your studio?
I was initially drawn to training my own custom AI image set because I wanted to find a way of working with AI that could still feel specific to my own practice and produce imagery that was less easily replicated. That led me to train a privately accessible LoRA model using only my own paintings as its visual archive. From there, I collage and overlay generated fragments in Photoshop, adding pixelated patterns with custom brushes and creating many layers with varying transparencies. Once I’m happy with the Photoshop file, I selectively erase areas to create interruptions and openings. I work with a fine art printing company to print the image directly onto cotton canvas with pigment. Once the canvas is stretched, I improvise with physical paint, responding to the digital base composition. I then photograph the painting and continue editing it on my iPad in Procreate, where I can introduce more distinctly digital, haptic gestures. I am especially interested in capturing the digital brushstrokes and traces that would be difficult to invent in paint alone. I then return to the physical canvas and try to translate those marks back into paint using a variety of techniques. The translation is never exact because of texture, gravity, and material resistance, but I continue moving between improvisation and planning until the painting feels resolved.
dawn jian, LightPulse, 2025. Acrylic, spray paint, and custom AI-generated digital image over archival pigment print on canvas, 60 × 22 in.
Q6. Scraping appears to be central to your practice, both as a physical gesture and as a conceptual echo of how AI systems scrape images and cultural memory from the internet. How did this connection between bodily gesture and algorithmic extraction develop in your work?
The connection developed somewhat unexpectedly. I was reading Empire of AI, which describes how early AI systems were trained by scraping enormous amounts of text and imagery from the internet. Around the same time, a gallery had looked through my work and chose to exhibit a painting I had made in 2019 called What Doesn’t Kill You (Makes You Stronger) that centered on the physical gesture of scraping black paint across a delicate pink ground. At the time, I had been experimenting with using black in my work, and scraping had allowed for some level of transparency to occur, revealing the glittering rosy colors underneath.
A Chinese brush painter who visited the exhibition commented on the linework in that painting. At first, I was puzzled because I did not think of the work as having many painted lines. She pointed out that the “lines” she was seeing were actually the clean edges left behind by my scraped marks. I really liked that observation. It made me think about scraping as a way of producing line indirectly, and as a way of lightly subverting the foundations of Chinese brush painting, where the expressive quality of the line is so central.
As I learned more about AI infrastructure, I became interested in the parallel between scraping as digital extraction and scraping as an aggressive gesture of removal. Continuing to scrape the paintings allows the physical process of making to echo the immense speed and scale of scraping used to train AI models. Bringing the two together allows the paintings to connect with algorithmic extraction in a more tactile way while also connecting back to my interest in Chinese brush painting through a more indirect and disruptive form of mark-making.
Q7. Many of your works move between digital signs, such as checkerboard-like transparency grids, and dense physical surfaces made through paint, scraping, and collage. In The Edge of Erosion, acrylic and metallic collage create a particularly layered, mineral-like surface. What draws you to this movement between digital image structures and tactile material presence?
That movement between digital and physical is central to my process. Each time an image passes between AI, Photoshop, Procreate, and paint, it changes. Clean digital marks become irregular through texture, gravity, and material resistance, and I’m interested in preserving those shifts in translation between the digital and physical rather than masking them away.
I’m also drawn to importing seemingly immaterial digital elements—checkerboard transparency grids, pixelation, and glitches—into painting and embedding them within a tactile canvas surface. Reading Glitch Feminism expanded the way I thought about the glitch, not simply as an error or failure, but as a rupture that can create space for transformation and resist fixed meanings around identity, technology, and the body. At the same time, I was reading about cyberpsychosis as a speculative condition in which machines seem to absorb or mirror the emotional turmoil and inner anxieties of humans. I became interested in bringing both possibilities into the work: the glitch as a space of liberation and transformation, but also as a sign of psychological instability, breakdown, or unease.
In The Edge of Erosion, metallic collage and layered acrylic build a surface that feels almost geological, like mineral deposits or sediment accumulating over time. I like contrasting that physical density with digital structures that suggest speed, compression, and instability. The work belongs to a broader series that considers AI not only as an image-making force, but also in relation to the ecological, political, and economic systems that make it possible.
dawn jian, The Edge of Erosion, 2026. Acrylic and metallic collage on canvas, 39 × 32 in.
dawn jian, LightStream, 2026. Acrylic, fluid acrylic, and metallic acrylic over custom AI-generated digital image on archival pigment print on paper, mounted on canvas, 36 × 36 in.
Q8. Can you tell us about your recent solo exhibition What the Earth Remembers? What was the starting point for the exhibition, and how did you approach the installation of the works in the space?
The starting point for What the Earth Remembers was a desire to return more deeply to the physicality of paint. After spending so much time working between AI, Photoshop, Procreate, and other digital tools, I wanted to return to painting as a slow process, a form of resistance to the attention economy. For this series, I wanted the act of making to remain highly visible through accumulation, scraping, erasure, and reconstruction, so that time, labor, and revision were embedded in the surface rather than concealed.
I was also very drawn to the landscape surrounding 西樵山美术馆 (Xiqiao Mountain Art Museum) and wanted the exhibition to feel connected to its environment. The first group of works takes cues from water, wind, and shifting light, while the later paintings move toward sedimentary rock and mineral strata, taking cues from the mountains. This transition from the surface into the ground also became a way of thinking more critically about AI as dependent on physical infrastructures, rare-earth minerals, energy, and global supply chains.
I approached the installation as a gradual movement from atmosphere toward density. The earlier works feel more open and fluid, while the later paintings become heavier, more layered, and more materially textured. I wanted the exhibition to move between landscape and infrastructure, digital interference and physical matter, so that nature and technology were not presented as opposites, but as increasingly entangled systems.
_JPG.jpg)
Installation view of dawn jian’s solo exhibition What the Earth Remembers, 2026, Xiqiao Mountain Art Museum.
dawn jian, Residual Drift, 2026. Acrylic and silver foil on canvas, 39 × 32 in.
dawn jian, Glitchfield, 2026. Acrylic and digital image on archival pigment print on canvas, 35 × 23 in.
Q9. You describe artificial intelligence not only as a method of image production, but as a system entangled with ecological extraction, political power, and economic infrastructure. How do you hold those larger conditions inside paintings that remain visually seductive, atmospheric, or even beautiful?
I think that is something I’m still trying to figure out. It is difficult to address systems as vast and complex as ecological extraction, political power, and technological infrastructure within a painting without reducing them to symbols or making the work feel overly didactic. I’m not sure painting can fully contain those conditions, but I’m interested in allowing them to remain present as underlying tensions.
Beauty and visual seduction are also extremely important to me because they create a point of entry into the work. I’m drawn to atmospheric color, delicate washes, and luminous surfaces, but I often interrupt those qualities through aggressive scraping, mosaic blurs, and areas of visual instability. Something that initially feels beautiful or inviting might also carry signs of erosion, interference, or disruption.
I think that contradiction reflects my own relationship to technology. AI can feel incredibly generative and exciting, while also being embedded within systems whose environmental, political, and economic consequences are often obscured or unevenly distributed. Rather than trying to resolve that contradiction, I’m more interested in allowing the paintings to hold attraction and unease at the same time.
Q10. The title Digital Gorge suggests both a landscape and a technological condition. What drew you to this title, and how do you see the work moving between natural space, digital space, and emotional atmosphere?
I was drawn to the title Digital Gorge because “gorge” implies a landscape formed through erosion—a space that has been carved out slowly through pressure, movement, and time. I liked placing that geological image next to the word “digital,” which suggests something flatter and more immaterial. The title holds those two conditions together without fully resolving them.
In the painting, I was thinking about landscape less as a literal place and more as an emotional and perceptual space. Atmospheric washes and organic forms can suggest water, rock, or distance, while pixelation and subtle glitches interrupt that sense of continuity. I’m interested in the moment when it becomes unclear whether you are looking into a landscape, a screen, or some hybrid space between the two.
There is also something psychological about a gorge to me—it can feel expansive and beautiful, but also frightening in its depth, like a rupture or a void. That ambiguity relates to what I think of as the digital sublime: a simultaneous sense of wonder, instability, and overwhelming vastness that can arise from our relationship with technology. I wanted the painting to hold that tension between natural and digital space without settling completely into either one.
dawn jian, Digital Gorge, 2025. Acrylic and spray paint on canvas, 26 × 39 in.
Q11. If you could train an AI model on something completely unexpected—not your paintings, but a personal archive, a smell, a childhood memory, a place, a dream, or something else—what would you choose?
I think I would train it on a sensory archive of places from my childhood in China, such as the smell of 鸡仔饼 (a traditional Cantonese savory-sweet biscuit), fragments of the sound of crickets at night, the humidity in the air, and other small sensations that are difficult to preserve in any concrete way. I’m interested in the gap between what we remember and what can actually be recorded.
Memory is already unreliable by nature. We tend to reconstruct the past rather than retrieve it exactly, and over time certain details sharpen while others disappear, merge, or become distorted. Sometimes a smell or sound can trigger an unexpectedly vivid memory, even when the visual image of that place has become vague.
I would be curious to see what an AI model might do with that kind of fragmented and subjective material. There is something strange about trying to convert something as private and intangible as memory into training data, especially when neither memory nor AI is particularly faithful to an original source. I think the model’s failures might actually be the most interesting part—the moments when it misunderstands, invents, or fills in what is missing. Those distortions could begin to resemble the way memory itself works.
Q12. Looking ahead, are there particular questions, materials, spaces, or forms of public encounter that you hope to explore more deeply in the next stage of your practice?
Looking ahead, I’m interested in gradually moving beyond the conventional painted canvas and thinking about how that structure might be deconstructed or reconsidered. I’ve recently started learning to weave, and I’m curious about eventually weaving some of my own painting surfaces, so that the canvas itself already carries pattern, structure, and labor before any paint is applied. I’m also developing a series of low-res ceramic 3D-printed objects that brings together L-systems and imagined ritual forms, almost as a kind of future archaeology.
More broadly, I want to keep exploring the tension between increasingly automated technologies and processes that are slow, tactile, and materially demanding. I’m interested in what happens when digital systems are translated through weaving, clay, or paint, and in continuing to think about the relationship between humans and machines as collaborators, each contributing different forms of decision-making, gesture, and unpredictability.
I’d also like to work more spatially with installation. I’ve been thinking about how an image might become an object, an environment, or something experienced through the body. I’m still figuring out what that could look like, but I’m interested in creating more immersive encounters where the process of making remains visible, perhaps integrated with sound or other sensory elements to build a larger atmosphere and make viewers more aware of their own physical presence within the work.