Pavel Kiselev
- Jan 26
- 5 min read
Pavel Kiselev is a Moscow-based artist who works with painting and new technologies, often starting from AI-generated sketches and then translating them into oil on canvas. He treats these digital images as raw material, selecting, cropping, and reworking them until they feel right to paint by hand. Titles usually come at the end, once the image has settled and a mood or moment becomes clear.He is interested in what happens when an image made by an algorithm is worked on by hand, and how paint can change something that started on a screen.
Q: What do you look for when you choose a digital sketch to turn into a painting?
A: When I work with digital sketches generated with AI, the selection process becomes a form of critical filtration. My goal is not to find a “successful image” in the conventional sense, but to identify what I call the visual nerve—an element that disrupts familiar perception and elicits an emotional response. The selection process consists of several stages. It all begins with generating a large set of sketches. This is similar to working with an archive or dataset. From hundreds or even thousands of options, I highlight those that have unusual structure, unconventional composition, or conceptual potential. It is also important to me that a sketch functions not only on its own but can also become part of a series—that it already contains the potential for a dialogue with other works.
Next, I test how the sketch works at different scales: projecting it onto a wall, enlarging or reducing it. Some images reveal themselves only in large format, while others work better when small. Sometimes I take only a part of the image—cropping the fragment that contains the main energy—and work with that.
This is crucial, because a digital sketch is not a finished composition; it is a material to interact with and transform. At the same time, I mentally translate the image into a painterly process: I imagine the thickness of the brushstroke, the texture of the surface, gloss or matte—or sometimes both—and possible techniques, from delicate glazes to impasto painting. If a sketch does not suggest a convincing material realization, it does not progress further. In the end, I choose the image where emotional impulse, compositional coherence, scale, and future materiality converge at a single point. This moment of convergence determines which digital sketch becomes the foundation for the painting.
Q: Moving from AI to oil on canvas, what changes once the image becomes physical?
A: For me, as an artist, the shift from a digital sketch to a physical painting is more than just a technical process. The digital sketch serves as a foundation, but when it’s brought to life on canvas, every brushstroke, every texture, and every detail gains its own unique life. This process involves unexpected adjustments, new decisions, and subtle nuances that set the physical artwork apart from its digital counterpart. Moreover, the physical painting offers the viewer a unique experience. When standing before it, the viewer perceives a tactile sensation through their gaze—feeling the depth, the texture, the energy of each stroke, as well as the interplay of light and color. This sensory engagement is something that digital formats simply cannot replicate. In essence, the physical painting is not just a reproduction; it’s a living, breathing entity that provides a richer, more profound connection with the viewer, which is unattainable in a purely digital realm.
Q: While creating, how much do you let the image lead you?
A: During the creative process, I strive to keep the initial concept as close as possible to the final outcome, especially during the sketching phase. However, as I delve into the work on texture, depth, and tonal nuances, the process inevitably becomes more intuitive and dynamic.
While the fundamental form and color scheme remain anchored to the original vision, the physical execution can lead to unexpected variations. This interplay between the planned and the spontaneous is what ultimately enriches the final artwork.
Q: “5 AM” and “At Dawn” both return to the moment between night and morning. Why that time?
A: I prefer to title my works only after the painting is complete, because during the process of contemplation and reflection, associations and sensations take shape that ultimately define the name. An important aspect is that the title and the image can exist independently. The image may evoke the viewer’s own associations, which do not necessarily align with my vision. However, the title helps guide the viewer’s attention, offering my perspective and inviting a dialogue. Thus, the viewer can discover both my interpretation and their own meanings, creating a unique experience of interacting with the artwork. For me, each of these titles carries a specific depth and personal experience. Within the series, I perceive these paintings as a pair. The title “5 AM” reflects the moment when the world is just beginning to awaken; in this time, I see the symbolism of the transition from darkness to light, from stillness to dynamics. In turn, the title “At Dawn” emphasizes this subtle transition when nature comes alive and forms begin to gain clarity. This liminal moment perfectly reflects the duality in my works, where abstraction and figurativeness intertwine, and where intuition and algorithms meet.
Q: In “Alphabet” and “Crane”, the works feel more like sign systems than scenes. How do you build an image without relying on figures?
A: These works engage with the concept of automatic writing, where the subconscious freely manifests through chaotic lines and forms. Initially, the images are created by an algorithm that generates abstract scenes based on set parameters; I then select and refine the sketches that most fully reflect my artistic vision. I title the paintings only after they are completed, relying on subjective intuition. I titled the piece “Alphabet” because I sense a distinct structure and system within it. The intertwining lines transform into forms, creating a structured space—much like letters combining into words, and words into sentences, carrying a hidden meaning revealed only to the beholder. “Crane”, conversely, reflects a free-flowing chaos. Here, the lines and forms lack a rigid figurative foundation, yet they still evoke associations with natural, organic shapes.
Q: Working with AI as part of your process, what do you want to push further next?
A: For me, artificial intelligence is essentially a tool—much like photography or digital editing software were in their time. It significantly accelerates and enriches the creative process. During my time in art school, the process of research and exploration was painstaking and time-consuming, involving extensive archival work and meticulous sketchbook documentation. Today, AI transforms this landscape, enabling rapid generation of visual material and streamlining the research phase. This not only saves time but also deepens the scope of creative exploration, allowing me to curate extensive databases of ideas and visual concepts efficiently. Moving forward, I aim to leverage AI to build a comprehensive repository of sketches and concepts, ultimately shaping a cohesive visual language. In essence, AI is not just a tool but a catalyst for deeper, more efficient creative development.
