What does “Chinese beauty” look like to the world today?
Is it red cheongsams, golden dragons, and paper lanterns — or something deeper, more emotional, more modern?
In my research project at Central Saint Martins, I set out to explore this question:
How might AI and aesthetics help reimagine China’s global image, beyond outdated symbols?
To do this, I combined cultural interviews, AI image generation, and visual testing with global users — creating a system of aesthetic archetype that represent the emotional and visual diversity of modern Chinese identities.
Step 1: Defining the Avatars
I began by observing real young people in China — friends, designers, students — and extracted five distinct aesthetic archetypes from their personalities, tastes, and daily rituals:
- Silent Grace – quiet, minimal, poetic
- Urban Pulse – edgy, energetic, streetwise
- Playful Spirit – colorful, chaotic, Gen Z confidence
- Rooted Flow – traditional, handmade, nature-connected
- Future Mist – cyber, cool-toned, rational
Each archetype is not a stereotype, but a felt experience. A way of being.
Step 2: Turning Words into Avatars
Using tools like Midjourney and GPT, I generated visual avatars for each archetype — not just clothing styles, but entire moods: lighting, space, body language, and color palettes.
These avatars were created with cultural prompting strategies inspired by the Kahani Project (Dwivedi et al., 2024), which showed that AI-generated visuals are more resonant when embedded with emotional and narrative cues.
Step 3: Testing the Impact
I showed these avatars to a diverse group of international peers and asked:
- Which one feels most Chinese to you?
- Which one emotionally draws you in?
- Does this change how you perceive Chinese beauty?
The results were both humbling and insightful.
Key Findings:
1. Perceptions were difficult to shift.
80% of respondents said the avatars did not change their perception of “Chinese beauty.”
Some cited past imagery from photographers like Chen Man as their reference point.
Without narrative or cultural context, most visuals were seen as “aesthetic,” but not “culturally meaningful.”
2. Emotion > Symbolism.
The most resonant image was not the one with the “most Chinese” look —
but the one with a smile, colorful tones, and a gentle caption.
People responded more to mood than to design details.
3. People crave context.
Several users said they didn’t fully “get” the images without story, explanation, or cultural cues.
This echoes research by Bell (2018) and Grix (2002), who emphasize the need for combining methods — narrative + visuals + testing — to build cultural understanding.
What This Taught Me:
Culture is not built from symbols alone.
It is built through emotion, context, and shared meaning.
AI can be a powerful tool — but only if we embed it with real voices, everyday beauty, and cultural nuance.
My project doesn’t offer a fixed definition of “Chinese beauty.”
Instead, it opens a door to ask:
What does it feel like to be modern, Chinese, and visible in the world today?
And how can we make that feeling seen — and felt — across cultures?
Leave a Reply