top of page
weidilogo.png
fig12.png

" In A Walled City, memory becomes spatial, identity becomes urban, and the city becomes an interface for collective speculation.

This is a cartography of lived data, built not from blueprints but from dreams, fragments, and the shared echoes of what once was—and what may yet be."

- Review by Auronda Scalera and Dr.Alfredo Cramerotti

A Walled City 2025

Weidi Zhang & Rodger Luo

What If our collective memories could shape a decentralized city?​

Can we learn from the historical complexity of a decentralized city to imagine new systems that structure our future virtual habitation,where memory becomes architecture, identity is spatialized,  and communities emerge through a rhizomatic network of collective presence?

A Walled City is an interactive AI installation that transforms personal and collective visual memories into a decentralized, evolving virtual city. Grounded in the spatial logic, zoning, and aesthetics of the historical Kowloon Walled City (Hongkong), a self-built hyper-dense enclave that grew in a legal and political void from the 1950s to 1993, the work builds a multi-agent AI system to turn participant-uploaded images into architectural chambers, forming a living structure where data, memory, and historical urban logic converge.
 

The project is not nostalgic; rather, it reflects on density, proximity, and how private fragments can be reassembled into shared space. Each contribution remains visible as itself, yet becomes part of a larger organism that accumulates, shifts, and evolves over time. In A Walled City, memory becomes architecture, identity is spatialized, and communities emerge through a rhizomatic network. The work connects AI systems with historical urban memory, turning contributions into forms and forms into relations, while raising the question of what is protected by “walls” and what is kept apart in the worldbuilding of future settlement.

Introducing  A Walled City ] 
[ 1 minutes teaser ]
Teaser | A Walled City, 2025
The Make of  A Walled City ] Research Documentary

[ 3 minutes documentation ]


This video provides an overview of the AI Art Research Project [ A Walled City ], an interactive AI art installation. It introduces the project’s core concept, historical grounding in the Kowloon Walled City, and the system design behind its multi-agent generative process. Through visual narration and excerpts from the installation, the video illustrates how individual visual memories are transformed into architectural chambers and collectively assembled into a continuously evolving, decentralized virtual city.
Research Documentary | A Walled City, 2025


----------------------------
All historical images and reference materials of Kowloon Walled City appearing in this video are used solely for illustrative and contextual purposes, in accordance with non-commercial, educational, and fair use principles. These historical materials are included ONLY to explain the historical background informing the project and are NOT affiliated with or representative of the artwork itself. Copyright remains with the respective rights holders or in the public domain. This video is shared privately for the purpose of presenting and discussing the concept and development of the artwork, and is not intended for public distribution or commercial use.
The Full Experience of  A Walled City ] Installation
[ 2 minutes documentation ]



This work is supported by

Herberger Institute Research-Building Investment 
Arizona State University

" A Walled City by Weidi Zhang and Rodger Luo is a generative monument to memory, density, and collective digital presence. Referencing the historical enigma of Hong Kong’s Kowloon Walled City—a space born of legal ambiguity and cultural resilience—the work reimagines the logics of decentralized habitation through a participatory, AI-driven installation. Participants contribute visual memories, which are transformed into digital chambers—each a mnemonic capsule—that assemble into an evolving urban tapestry.

This living system, a bespoke multi-agent AI, trained on the spatial and social intricacies of Kowloon, autonomously generates new architectural volumes based on the submitted images. Layers of computational interpretation turn photos into speculative micro-architectures, governed by synthetic agents with distinct urban roles—constructing grocery peripheries, rooftop laundries, nested corridors of memory. Projected as an immersive environment, the installation allows viewers to witness their own memories woven into an algorithmic cityscape—one shaped not by central planning, but by the entanglement of human input and machine agency. In A Walled City, memory becomes spatial, identity becomes urban, and the city becomes an interface for collective speculation. This is a cartography of lived data, built not from blueprints but from dreams, fragments, and the shared echoes of what once was—and what may yet be."

- Review by Auronda Scalera and Dr.Alfredo Cramerotti

out1.jpeg

EXHIBITION FORMAT 

[ INTERACTIVE AI ART INSTALLATION ] OR [ AUDIO / VISUAL PERFORMANCE ]

 

AWARD:

2026      [ WINNER ] IF DESIGN AWARD GERMANY

2025      [ GOLD WINNER ] MUSE DESIGN AWARD UNITED STATES

2025      [ GOLD Tickets ] AI Art Mag, Selected by Curators Auronda Scalera and Dr. Alfredo Cramerotti

 

EXHIBITIONS:

2026      [ INTERACTIVE VERSION 2.0] UPCOMING - International Symposium For Electronic Arts (ISEA), Dubai 

2025      [ INTERACTIVE VERSION 2.0] MAD Arts Museum, Florida, US 

2025      [ INTERACTIVE VERSION ] IEEE VIS Art Program, Vienna, Austria 

2025      [ INTERACTIVE VERSION ] Center For Digital Media, Signal [ VIFF x DigiBC ], Vancouver, CA

2025      [ INTERACTIVE VERSION ] Media and Immersive eXperience Center, Arizona State University, Mesa, AZ, US

2025.     [ INTERACTIVE VERSION ] CVPR AI Art Gallery, US

2024.     [ AUDIO/VISUAL PERFORMANCE ] Lumiere Hall, Metaxis Festival, Budapest, Hungary

PRESS:

Review by Auronda Scalera and Dr. Alfredo Cramerotti, AI Art Magazine, 2025

Zhang, W. (2024, October). A Walled City. PROMPT MAG Book, (11), 49.

fig7.jpg
MAD-installation-view-.jpg
bottom of page