Bowen Shen · CLIENT: Bowen Shen
This generative art series depicts everyday childhood moments through AI assisted image making. The work focuses on scenes such as school recess, quiet summer afternoons at home, and New Year celebrations, drawn from lived experience in Northern China. Created through guided prompting and manual visual editing, the series prioritizes realism while allowing subtle visual irregularities that reflect how memory preserves feeling rather than exact detail.
Press Kit
- Unique Properties
- As I Remember Them is a generative art series that visualizes fragments of memory, emotion, and longing. It is an attempt to reconstruct the past through the lens of imagination.
- Dimensions & Packaging
- 1856 x 2464px image series
- Tags
- Generative AI art, Nostalgia, Childhood memories, Northern China, Visual storytelling, MidJourney, Photoshop, Digital art
- Production Technology
- This series was created with MidJourney, using iterative prompting and reference-based descriptions to shape culturally specific scenes. Outputs were then refined in Photoshop through compositing, retouching, and color adjustments, ensuring the final work aligned with my artistic vision and emotional intent.
- Design Challenge
- The greatest challenge was translating personal, culturally specific memories into AI-generated images. MidJourney lacked context for the textures of Northern Chinese life, requiring me to break scenes into elements, source references, and iteratively refine prompts. Balancing realism with the quirks of AI output, and deciding when to embrace its distortions, was both the obstacle and the creative breakthrough.
- Project Duration
- The project was created virtually over the span of one month in January 2025.
- Operation Flow
- The generative art is experienced through sequential viewing. Each image presents a self-contained moment, allowing viewers to pause, observe, and form personal associations. The series encourages slow viewing rather than interaction through controls, inviting audiences to connect individual scenes to their own memories and emotional experiences.
- Research
- Process documentation: https://bit.ly/49OvkvG
Type of Research: Practice-based creative research.
Research Objectives: To explore how AI can represent personal and culturally specific memories.
Methodology: Iterative visual experimentation and technical study.
Data Collection and Tools: Memory recall, curated reference images, MidJourney style parameters, prompting techniques, Photoshop.
Participants or Experiments: Repeated prompt and image generation cycles.
Results: Increased control over style, detail, and emotional accuracy.
Insights and Impacts: Structured prompting reduces emotional flatness in AI art.
Real Life Effects: Encourages more human-centered AI use in design and visual storytelling. - Inspiration
- Bothered by how often GenAI outputs felt soulless, the project set out to render a deeply soulful subject, the emotional landscape of childhood.
I grew up in Northern China, in a hometown changed by time and distance. Through prompting, storytelling, and hands-on editing, I recreated scenes that linger in me: recess, summer solitude, New Year warmth, while embracing GenAI quirks that reflect how memory distorts and softens.
Press releases are available in 1 languages.