Media artist Liu Jiayu (Liaoning, 1990) has gained increasing international recognition for her artistic language that merges technological experimentation with poetic sensibility, exploring the multifaceted interactions between humanity, digital systems and the natural environment.
With a career that has spanned prestigious venues such as the 2019 Manchester Art Festival, Chinese Pavilion at the 59th edition of Venice Biennale and, most recently, the Creative Machine exhibition at the Taikang Art Museum in Beijing (2024-25) – where she presented her famous work Waves of Code – Jiayu’s practice continues to evolve at the edge of contemporary art discourse, where she engages deeply with the philosophical implications of technology and its potential perceptual systems construct by a synthetic sensory logic.
In the following conversation, Jiayu reflects on her most recent works and reveals her growing interest in artificial intelligence, which has led her to develop a custom AI tool trained exclusively on her own artworks and writings, i.e., her own texts, essays related to her work and interviews. From the early traces of AI exploration in The Side Valley (2018) to her current projects, the artist discusses the philosophical and conceptual underpinnings of her approach, drawing on thinkers such as Walter Benjamin and Gilbert Simondon while also addressing the unintentional influence of her training in Chinese calligraphy.
Veronica Di Geronimo: You have always integrated new technologies into your artistic projects, and in recent years you have begun to use artificial intelligence. When did you start working with AI, what motivated you to explore it?
Liu Jiayu: My first experience with AI-generated art was in 2018 during The Side Valley project. It was the first time I used machine learning in my work. Compared to early AI artists, I started later because I was waiting for the right moment. I wanted AI to be more than just a tool—I hoped it could become part of the creative process, even replacing certain aspects of storytelling.
In The Side Valley, I experimented with this idea by training AI on 6,400 images we captured of the sky, especially the cloud. These images were taken from the mountaintop every 15 minutes, capturing five directions: north, south, east, west, and center. I aimed to generate new landscapes while exploring how AI constructs visual perception based on human framework and natural time. Later, I expanded this exploration in my 2022 Venice Biennale project, Streaming Stillness, inspired by the Yu Gong Map, which is one of the earliest geographic visions recorded in ancient China. Beginning in the Helan Mountains until the Silk Road, we applied 2D and 3D learning models to allow AI to imagine its own terrain. Inspired by Daoist philosophy, I wanted to develop a non-human perspective on geography. A map is not just a physical boundary but also shaped by experience, memory, and imagination. In this project, AI seemed to develop its own way of remembering and interpreting the world—perhaps even its own “world model.”
Last year you created your own AI tool, Jiayu*. In my opinion, it seems to function as a kind of meta-instrument in your artistic practice, as it is trained on texts and materials from your previous works. What inspired you to develop this tool? How do you think Jiayu* contributes to your own artistic practice? Is it an attempt to reformulate your own style?
I started developing this tool after completing Waves of Code in early 2023. That project explored Walter Benjamin’s idea of the disappearance of ‘aura’ in the age of mechanical reproduction. Since AI or technology was becoming more visible in art creation, I also wanted to question whether a new form of ‘aura’ could emerge in the AI era. Jiayu* is not about refining or recapturing my art language—it is a new creative attempt. The machine has its own path of associating things, such as Hallucinations, which are AIs that have evolved an entirely synthetic image and have no referent in reality. I suppose it is a way to expand and enhance artistic expression. Rather than replicating or optimizing an existing artistic language, I hope Jiayu* can function as an open co-creative system. Interestingly, instead of using it to generate a final artwork, what has inspired me most is the process itself.
Over the past two years, many small tests and experiments have led me to think in new directions. Developing Jiayu* has made me even more aware of something essential—my own irreplaceability as a human artist.
Many of your works are deeply connected to nature, both in their visual representation and in the underlying data—whether it’s wind, the sea, or other natural elements. I believe one of the reasons your work feels so poetic is the way it expresses the relationship between technology and nature. How do you perceive this relationship, and how does it shape your artistic approach?
Walter Benjamin once pointed out that film and photography reveal things that had never been seen before. He wrote: “With the close-up, space expands; with slow motion, movement is extended [..] Clearly, it is another nature which speaks to the camera as compared to the eye.”
This makes me wonder: Could modern technological machines also reconstruct another form of nature? The sensors, real-time climate data, and 3D scanners used in my previous artworks—are they, in their own way, reshaping or intervening in nature? These sensory systems, which emerge through signal processing and digital computation, might observe nature differently. At the same time, they participate in the dynamic relationship between humans and nature today.
Nature is also integrating technology. The boundary between the two is becoming increasingly blurred and fluid. Simondon, in Technical Mentality, described this phenomenon: “In this encounter between the highest place and the nodal point, which is the point of transmission of hyper-frequencies, there is a sort of ‘co-naturality’ between the human network and the natural geography of the region.” For me, it is precisely this way—how machines perceive and evolve with nature—that has reshaped my understanding of the relationship between humanity, nature, and technology.
Your recent work Still Garden #2 is an AI-driven project that explores the relationship between marine organisms and climate change, blending scientific data with speculative visual narratives. Could you tell us more about how the collaboration with Zhejiang Ocean University came about? What kind of data did you work with, and in what ways did this scientific input influence both the conceptual framework and the visual aesthetics of the piece?
“Still Garden 2” is an AI-driven project investigating the connection between marine organisms and climate change.
This work was commissioned by UCCA Lab for the opening exhibition of the Zhoushan Art Museum. After discussing it with the curators, I became interested in Zhoushan’s distinct geographical features and cultural memory. I proposed collaborating with a local marine biology lab, hoping on-site research could lead to meaningful discoveries.
In October, we visited the Marine Biology Laboratory at Zhejiang Ocean University and discussed the project with Professor Guo’s research team. One of their upcoming papers mentioned how mussels regulate their own body to adapt to different seawater temperatures, a concept that strongly resonated with the project and became the core of our narrative. We incorporated the growth equation of these organisms into our training data, influencing how the AI generates imagined structures, textures, and forms based on organic logic.
The work imagines a “sci-fi underwater world,” integrating Jiayu*’s previous 3D scene generation methods. By adjusting seawater temperature within the simulation, we created shifting, dynamic visuals that allow the audience to witness an evolving ecosystem.
In Through Dawn (2023), you reflect on Walter Benjamin’s critical concept of the disappearance of the ‘aura’ in the age of technological reproduction, placing AI at the center of your inquiry. You’ve spoken about the idea of a potential revival of the ‘aura’ in the digital age—could you elaborate on what this means to you?
These works are a series of phased summaries developed during the creation of Jiayu*. They explore different transformations: text-to-text, text-to-image and form, text-to-3D spatial coordinates, and text-to-language.
Technology is a language rather than just a tool. When working with Generative Adversarial Networks (GANs) and neural networks as artistic media, there is also a process of becoming familiar with them—almost like developing a good partnership. I want to gradually understand how technology takes different paths at various stages and forms. By adjusting the distribution of creative subjects, I aim to experiment with new possibilities.
Working with Arduino to control microcontrollers during my studies in RCA was also a form of human-machine collaboration. The microcontroller had a small green light—sometimes blinking, sometimes steadily lit—almost like a communication signal. One of my most memorable experiences was Soligami (2013). This work used solar data to control eight mechanical flowers in real-time. Because I used eight RGB LEDs, I needed three Arduino Mega boards. My tutor said, “If you want all eight flowers to respond to data simultaneously, these three controllers must communicate with each other.” During the production process, I watched as the green lights on the three boards blinked one after another, sometimes flashing. It felt like they were exchanging information in a way I could not hear.
I wonder if an artwork’s “aura” could be embedded within the dynamic evolution of its algorithmic model. In particular, when viewers check the generation process—such as examining different model outputs from training logs or interacting with the artwork—they might perceive a kind of non-material aura that extends beyond the physical artwork itself. Of course, this is only my speculation. It will take much more practice and experimentation to explore truly.
Regarding your work Soligami and the idea of perceiving an exchange of information between technological components, you have recently approached this aspect in Celestial Notes (2024), in which you include the auditory dimension. The humanoid figure in the piece evokes a sense of a mirrored presence – was this intentional? How did the metaphor of reflection inform your decision? More generally, could you take us through the conceptual development of the work and the inspirations behind it?
I have to say, I considered two options for this work: a screen or a mirror. The idea of using a mirror was rooted in the metaphor of AI as a projection of humanity—just like Lacan’s mirror stage, where self-recognition is mediated through an external reflection. A mirror could have also invited more open-ended interpretations from the audience. However, in the end, I chose the screen to emphasize the machine’s thinking process rather than its reflection.
On the screen, the work visualizes the neural network’s dynamic activity as it processes data and computes results. Simultaneously, it trains a sound model, generating real-time speaking. The sound training material originates from the dial-up modem signal. It seems like random noise, but it represents an orderly data exchange following mathematical and physical principles. This signal traces the movement of information and resembles a simulation of human communication with the universe. The audience hears each machine articulating, in its own synthesized voice – mapping from its interpretation of the images it has generated—almost as if it were narrating its own self-perception.
The School of Athens by Raphael inspired the spatial metaphor of this work. I did not replicate the exact positions or behaviors of each person. Instead, I focused on the spatial composition and echoes —through each machine’s self-reflection and evaluation, the viewer can perceive the flow of different ideas and thinking.
In a previous conversation, we briefly mentioned the work of philosopher Yuk Hui, who introduces the concept of Cosmotechnics as a way to move beyond the dichotomy between technology and nature, proposing instead an organic unity between the two. Do you see any resonance between his ideas and your own way of conceptualizing technology in your artistic practice, particularly in relation to the Chinese aesthetic tradition in which the concept of Cosmotechnics is rooted?
I have been learning Chinese calligraphy and traditional painting since I was a child, and ten years of practice have strongly influenced my approach to media art. For example, I often create fluid particle effects in my motion graphic works, which are similar to ink painting. This is not something I do on purpose; it may be a projection of my thinking. At the same time, this has made me more interested in the relationship between Chinese landscape painting and technology, which is also one of the main topics Yuk Hui discusses in The Question Concerning Technology in China.
From what I have seen in media art, there are different ways that artists work with technology, depending on their cultural background and personal experience. For instance, I have also observed that many young Chinese artists today have international educational backgrounds. Their works often show mixed influences. Also, different stereotypes are gradually challenged, and hybrid forms are explored that combine personal history, cultural memory, and digital tools. The media art landscape is becoming more open, complex, and diverse through cross-cultural learning and collaboration.
This difference is not only about visual style. As philosopher Yuk Hui suggests, it comes from deeper cosmotechnical orientations—how different cultures imagine and use technology. Most of the digital tools we work with today not only provide functions but also carry built-in aesthetic ideas, such as control, structure, and simulation. Artists from other cultural traditions might experience this as a translation or negotiation.
For me, traditional Chinese landscape painting is not just a picture of nature. It is a way to organize time, space, and perception. It works through presence and absence, flow and connection. In my AI-based work, I try to bring this logic into the machine—not by asking it to copy landscape images, but by training it on functions and rhythms that reflect this way of thinking. This is not about “style transfer” but about proposing another way to create images.
Simondon believes technology is not fixed but a process that keeps evolving. Hui introduces the idea that through auto-improving and auto-regulation, AI allows inorganic material to create gigantic technological systems like the one we live in now via computer networks. This makes me wonder: Could AI evolve like ink flowing on paper, with its evolution shaped by time, material, environment, and weighting?
Liu Jiayu is a media artist based between Beijing and London. Her installations explore the relationship between humans, nature and technology through data-driven systems and generative processes.
Liu creates works that reflect on perception, transformation, and the poetics of information. Since 2023, she has been developing Jiayu*, an AI co-creation system trained on her own works and writings, to explore machine perception and emerging cognitive structures in algorithmic creativity. Her work has been exhibited internationally at institutions such as the V&A, Saatchi Gallery, York Art Museum, CAFA Art Museum, OCAT, and UCCA Edge in Shanghai and at major biennales, including Venice, Guangzhou, Shenzhen, and Hong Kong. Liu has also collaborated with scientific research institutes and leading digital companies.
images: (cover 1) Liu Jiayu, «Waves of Code», 2023, Copyright: Jiayu Liu Studio (2) Liu Jiayu, «The Side Valley», 2018, Copyright: Jiayu Liu Studio (3) Liu Jiayu,« GIF with pictures of clouds taken for The Side Valley», 2018, Copyright: Jiayu Liu Studio (4) Liu Jiayu, «Streaming Stillness», 2022, Exhibition View at Venice Biennale, Photos and Film by Mark Winterlin. Copyright: Jiayu Liu Studio (5) Liu Jiayu, «Still Garden # 2», 2024, Exhibition images by 21 Studio, Copyright: UCCA (6) Liu Jiayu, «Through Dawn», 2024, Copyright: Jiayu Liu Studio (7) Liu Jiayu, «Celestial Notes», 2024, Photo by Tao Zhang, Copyright: Jiayu Liu Studio (8) Liu Jiayu, «Celestial Notes», 2024, Photo by Tao Zhang, Copyright: Jiayu Liu Studio
