Software engineer and creative technologist Jing Dong has released SoulArts Art & Dreams, an artificial-intelligence application that turns drawings, dreams, and written reflections into material for creative self-exploration.
Developed through inno Art & Tech Studio, the interdisciplinary studio she founded, SoulArts marks a new stage in Jing’s work at the intersection of artificial intelligence, interaction design, and contemporary art. Together with projects such as the experimental Hearing Space-Time, it extends questions she has investigated for years through interactive installations, performance, wearable devices, and augmented reality.
“I am interested in what happens when something deeply human - an emotion, a gesture, a memory, or a dream - enters a computational system,” Jing said. “The goal is not to let the machine define that experience, but to create another way for a person to encounter it.”
From emotional data to generative images
In 2024, Jing presented Emotion Capture Station (2020) in Porous Constructions, a collaborative group exhibition organized by CAAN and Unveil Gallery and curated by Moham (Zhang) Wang. The exhibition brought together painting, sculpture, installation, photography, and video to examine “porosity” in visual and curatorial practice.
Jing’s work translated human emotional data into generative visual art. Rather than treating emotion as a stable category, the system asked what might be changed, emphasized, or lost when an internal state is processed by an algorithm and returned as an image.
The project introduced a question that now runs through much of Jing’s technical practice: When emotion meets an algorithm, how is it understood - and who has the authority to interpret the result?
For Jing Dong, a computational output is not a final answer. It is a new layer of material that a person can examine, question, or reject. The system is therefore designed around an interpretive loop: Human input enters the machine, the machine transforms it, and the transformed result returns to the person as the starting point for another response.
An engineering career reoriented toward human experience
Jing Dong studied computer science before working for several years as a front-end software engineer. Her professional background includes software engineering experience at Instagram, where working with consumer-facing systems gave her a close view of how interfaces shape attention, behavior, and communication.
In commercial technology, software is usually built around a predetermined function: Complete a task, guide a user toward an action, improve efficiency, or produce a measurable result. Jing Dong became interested in a different kind of technical problem - how to build systems that could remain open-ended and support reflection rather than prescribe a response.
She entered the Art and Technology program at the California Institute of the Arts in 2019 and received her master’s degree in 2021, while also studying in the school’s Integrated Media program. The move did not represent a departure from engineering. Instead, it allowed her to redirect software engineering toward questions involving perception, embodiment, emotion, and machine interpretation.
“Code is often treated as invisible infrastructure,” Jing said. “But every system contains decisions about what can be perceived, what can be measured, and what kind of response is possible. I want to make those decisions part of the experience.”
Building a feedback loop between movement and language
In the 2021 new-media performance Poetic Imagination X AI, Jing designed a two-part artificial-intelligence system that translated a dancer’s movements into poetic language.
A movement-recognition component interpreted gestures as keywords. A language-generation component developed through a GPT-2 framework then transformed those keywords into lines of poetry. The dancer encountered the machine-generated text and responded with a new movement, which returned to the system as another input.
The performance operated as a recursive exchange among body, data, and language. Choreography no longer moved in a single direction. It circulated: Movement became text, text prompted movement, and each participant became the other’s point of departure.
Jing expanded this approach in The Hidden Space (2022), connecting a wearable sensing device with cloud-based language processing. The device collected sounds from a performer’s surroundings - including fragments of speech, ambient noise, and accidental interruption - and the software transformed that material into poetic text in real time.
These projects were built in different media, but shared a common system architecture: Sensing, interpretation, transformation, and human response. They also treated computational uncertainty not as an error to be eliminated, but as a condition that could produce new associations.
Between Physical Matter and Virtual Space
Although software forms the technical foundation of Jing practice, many of her earlier works examine the weight, fragility, and specificity of physical objects.
In Material Sound (2022), translucent CD cases were arranged across a deep-blue surface. Inside them, Jing placed tree roots, fallen petals, and other small remnants gathered from everyday life. Each object carried a texture, history, and form that could not be perfectly reproduced. Their material presence created a counterpoint to the nearly weightless computational processes running through her other projects.
In Virtual Sculpture (2022), the object became less stable. A bright, irregular form appeared through augmented reality in a desert landscape and on an urban street. It seemed monumental, yet it could not be touched. The work asked how a virtual object can alter the perception of a physical location even when it has no material body of its own.
For Jing, the digital and physical are not separate worlds. A generated image can change how a room is experienced; an invisible algorithm can influence the movement of a body; and an ordinary fragment of nature can interrupt the apparent smoothness of a technological system.
SoulArts: Designing an “emotional interface”
Dong’s latest application, SoulArts Art & Dreams, brings that research into a mobile product designed for individual creative exploration. Released for iPhone through Apple’s App Store, SoulArts combines free drawing, AI-generated imagery, dream recording, a personal diary, guided reflection, audio, and augmented reality.
The application begins with the idea that some experiences become visible before they become verbal. Anxiety, loneliness, memory, and other difficult-to-name emotions may first appear through a color, repeated symbol, image, or spatial arrangement. SoulArts gives users ways to work with this visual material without requiring formal drawing skills.
Users can sketch freely and invite “Soul AI” to transform a drawing into a more developed expressive image. Guided sessions combine creative prompts, reflective questions, audio, and music. Users can also record dreams, receive AI-generated reflections on symbols and patterns, revisit entries in a private dream journal, and write diary entries with optional AI assistance. An augmented-reality feature allows finished artwork to be placed within the user’s physical surroundings.
Jing describes SoulArts as an “emotional interface” - a digital environment in which people can use images, words, dreams, and sound to explore internal experience.
“SoulArts is not designed to tell users what their images mean,” she said. “It offers another image, association, or question. The user remains the person who decides what is meaningful.”
This principle distinguishes the application from systems designed to classify a person’s emotional state or provide a clinical diagnosis. Jing frames SoulArts as a creative self-reflection tool, not a medical or therapeutic service. Its interaction model is nonjudgmental: The AI does not score the user’s work or provide a standard interpretation.
Instead, the application creates a repeated exchange. A person makes an image; the system generates a visual extension; the person examines the relationship between the two and chooses how to continue. The process turns AI from a content-generation shortcut into a reflective collaborator.
From mobile software to a physical art ecosystem
SoulArts is one of several projects being developed through inno Art & Tech Studio, which Jing founded as a platform for work involving AI-assisted creativity, interaction design, augmented reality, animation, and spatial emotional intelligence. Operating between Los Angeles and Beijing, the studio brings together software engineering, product design, and artistic research.
The studio is also working to extend SoulArts beyond the phone. Jing said the team is developing options for users to order physical prints of artwork created in the application. A related electronic-frame project would allow changing digital work to be displayed as an object within a home or other physical environment.
The move from screen to object continues a concern visible throughout Jing’s earlier installations. Digital images may be generated remotely and remain physically intangible, but they still alter the spaces in which people live. Printing and displaying the work introduces material scale, texture, duration, and domestic context into an experience that begins with software.
Other projects incubated by the studio include Elumera and Hearing Space-Time, an experimental application released through Apple’s App Store. Hearing Space-Time uses machine vision to infer spatial depth and structure from an image, then maps those relationships into interactive sound. By touching different parts of the image, a user can listen to how the computational model has reorganized visual space. The compact experiment reflects a recurring concern in Jing Dong’s work: Making an otherwise invisible process of machine perception available to human senses.
Technology as a participant, not an answer
Across Jing Dong’s installations and applications, software performs different translations: Gesture into language, sound into poetry, emotion into image, and visual space into sound. These transformations connect works that might otherwise appear to belong to separate fields.
They also complicate the familiar question of whether artificial intelligence can be creative. Jing does not present AI as a replacement for the artist or user. Human experience, in her view, emerges from the particularity of memory, consciousness, embodiment, and personal history. At the same time, she does not treat the machine as a neutral instrument. Models introduce patterns, omissions, errors, and unexpected results that can reshape the direction of a creative process.
“The viewer - or, in the case of an application, the user - is the person who ultimately completes the work,” Jing said. “The system can generate a possibility, but meaning still depends on human participation.”
Jing’s projects have appeared in exhibitions and programs associated with Ars Electronica, JACCC’s Release: Heard in LA, Coaxial Arts’ Art + Technology programming, and Unveil Gallery. With SoulArts and Hearing Space-Time, she is now translating that exhibition-based research into publicly available software and developing a studio structure capable of supporting future interactive systems.
From emotional data to generative imagery, and from machine vision to spatial sound, her work continues to return to one central question:
How can technology help people perceive what they have not yet found a way to express?
“The most valuable technology may not be the technology that gives us the fastest answer,” Jing said. “It may be the technology that helps us ask a more meaningful question.”
This article was written in cooperation with Jing