Realm of Floating Silence, Chun-Chiao Wu on Regenerating Sacred Spaces Through AI Collaboration
Open Access Peer Reviewed Research Offers Institutions a Framework for Transforming Abandoned Sacred Sites into Vessels of Cultural Rebirth
TL;DR
Designer Chun-Chiao Wu created a peer-reviewed methodology for regenerating abandoned chapels using AI as a creative partner. The framework combines hand sketches, generative AI, and physical modeling. Cultural institutions now have a documented, replicable approach for transforming silent heritage sites into meaningful spaces.
Key Takeaways
- A hybrid workflow combining hand sketches, generative AI, and physical modeling creates meaningful spatial narratives from abandoned heritage sites
- Three analytical lenses covering emotional, spatial, and ecological dimensions provide systematic structure for sacred space regeneration
- Positioning AI as collaborative partner rather than automation tool maintains human creative authorship while expanding design possibilities
What happens when a childhood chapel falls silent?
Across the world, sacred spaces stand empty. Communities migrate, congregations disperse, and the buildings that once held generations of memory begin their slow return to the earth. For governments managing cultural heritage portfolios, for universities studying architectural preservation, and for institutions tasked with stewarding historical sites, abandoned structures present a fascinating question: How do you breathe new life into a space defined by silence?
Abandoned sacred spaces and their potential for renewal represent precisely the terrain that designer and researcher Chun-Chiao Wu explores in a peer-reviewed study that has captured attention across the design and architecture community. The research, titled "Poetic Regeneration of a Childhood Chapel through AI-Human Collaboration: A Design Inquiry into the Realm of Floating Silence," presents something rather unexpected. Wu's study offers a documented framework for how artificial intelligence can serve as a creative collaborator in transforming ruins into vessels of cultural renewal.
The work centers on a specific site: a chapel that shaped Wu's childhood, now left in decay as local communities moved away. Rather than approaching decay as a problem requiring a technical solution, the research frames the abandoned chapel as poetic ground for imaginative rebirth. The methodology combines hand-drawn sketches, generative AI tools, physical modeling, and exhibition presentation into a hybrid workflow that other institutions can study and adapt.
What emerges is a compelling case study for how organizations might approach their own abandoned sacred sites. The research has already demonstrated public resonance through presentation at the Taiwan International Interior Design Expo. Wu's findings offer concrete insights for cultural ministries, heritage foundations, architecture programs, and design studios seeking new approaches to spaces where traditional preservation methods may require supplementation with fresh creative perspectives.
The Phenomenon of Abandoned Sacred Spaces and Their Cultural Weight
Consider the scope of the abandoned sacred space phenomenon for a moment. Sacred spaces occupy a unique position in the built environment. Churches, temples, chapels, shrines, and ceremonial halls carry layers of meaning that extend far beyond their architectural footprint. These buildings witnessed births and deaths, celebrations and mourning, individual prayers and collective rituals. When sacred spaces fall into disuse, something more than a building becomes abandoned. A container of collective memory sits empty.
For institutions responsible for cultural heritage, the phenomenon of abandoned sacred sites presents both an opportunity and a responsibility. Government heritage departments often manage inventories of sacred sites without clear frameworks for intervention. Universities with architecture and design programs seek compelling case studies that bridge technical skill with humanistic concern. Cultural foundations look for approaches that honor the spiritual dimension of space while remaining practical and executable.
The traditional options tend toward either full restoration or managed decline. Full restoration preserves the original form but may freeze a space in time, disconnected from contemporary community needs. Managed decline documents and stabilizes while accepting eventual loss. Wu's research suggests a third path: regeneration that honors memory while creating something genuinely new.
The specific chapel at the center of Wu's research was once a place of gathering and growth. Local rituals brought community members together within the chapel's walls. Then, as with so many rural and semi-urban sacred sites, the population shifted. Young people moved to cities. The congregation thinned. The building stood increasingly quiet until silence became the chapel's defining characteristic.
The trajectory of declining congregations and increasing silence will sound familiar to anyone who has worked with heritage properties. The research takes the common narrative of sacred site abandonment and transforms the narrative into a springboard for methodological innovation. The question shifts from "How do we prevent decay?" to "How do we compose with decay, silence, and memory as creative materials?"
Understanding the Hybrid Human-AI Workflow Methodology
The methodology Wu presents deserves careful attention from institutions considering how to integrate emerging technologies into their design and preservation practices. The workflow proceeds through distinct phases, each involving a specific interplay between human intuition and algorithmic capacity.
The process initiates with field observations and what the research terms "emotional mapping" of the site's decay. The field observation phase is entirely human. The designer walks the space, notes the qualities of light falling through damaged roofing, observes how vegetation has begun reclaiming structural elements, and registers the atmospheric residue of past use. The observations become documented through narrative writing and hand-drawn sketches.
The second phase introduces generative AI tools. The research employed Stable Diffusion and ControlNet as primary instruments. The methodology becomes particularly instructive for institutions at the generative AI phase: the AI-generated images are explicitly treated as conceptual collaborators rather than finished products. The designer uses carefully engineered prompts aligned with hand-drawn spatial visions to generate imagery that extends the poetic and atmospheric qualities of memory.
The distinction between finished products and conceptual collaborators matters significantly. Many organizations approach AI image generation with an expectation of finished outputs. Wu's methodology positions generative tools differently. The AI-generated images become part of an iterative conversation, offering unexpected interpretations of prompts that the human designer can accept, reject, or modify. The research describes the collaborative process as "recomposing memory, material, and atmosphere" through human-machine dialogue.
The third phase moves into physical and digital modeling, translating the conceptual explorations into spatial compositions. Layers of petal-like translucent forms emerged from the modeling process, along with studies of light behavior and spatial rhythm. The methodology maintains connection to the original site while allowing for speculative departure.
The fourth phase involves in-situ exhibition, presenting the work to public audiences who can respond to and validate the approach. The presentation at the Taiwan International Interior Design Expo provided validation, with strong audience resonance affirming the methodology's capacity to generate meaningful spatial narratives.
For universities developing AI-integrated design curricula, the hybrid workflow offers a documented pedagogical model. For design studios seeking to incorporate AI tools without surrendering creative authorship, Wu's approach demonstrates a balanced method. For government heritage departments, the methodology suggests possibilities for community engagement through speculative design processes.
Memory as Design Material: The Poetic Dimension
One of the most valuable contributions of Wu's research lies in the articulation of how memory, emotion, and atmosphere can function as legitimate design inputs within a rigorous methodology. The articulation of memory as design material moves beyond purely technical considerations into territory that cultural institutions often struggle to address systematically.
The research frames the abandoned chapel through three lenses: emotional, spatial, and ecological. Each lens offers different information about the site and the chapel's potential for regeneration.
The emotional lens captures what the space meant to the chapel's users and what residue of meaning persists in the current state of decay. For the designer, personal childhood memories of the chapel provided the emotional foundation. For institutions working with abandoned sacred sites, analogous sources might include oral histories from former congregants, archival photographs documenting past gatherings, or community engagement sessions that surface collective memories.
The spatial lens analyzes the physical configuration of the site, including proportions, sight lines, acoustic properties, and relationship to surrounding landscape. In states of decay, spatial qualities often become more pronounced rather than diminished. Walls with openings to the sky create new relationships between interior and exterior. Floors returning to soil introduce texture and organic patterning.
The ecological lens considers how the site has been reoccupied by natural processes. Vegetation growth, animal habitation, weathering patterns, and material decomposition all become part of the design context. The research suggests natural processes can be incorporated rather than reversed.
What makes the three-lens framework particularly useful for institutional adoption is the framework's systematic quality. The emotional, spatial, and ecological lenses provide a replicable structure for analyzing abandoned sites while remaining open to the specific qualities of each location. A heritage department could train staff in the analytical approach. An architecture program could assign students to apply the framework to local sites. A design studio could use the three lenses as a client-facing methodology for heritage commissions.
The poetic dimension emerges from the synthesis of the three lenses rather than from any single analytical perspective. When emotional memory, spatial configuration, and ecological process are composed together, the result carries meaning that exceeds purely functional considerations.
Practical Workflow: From Sketch to Exhibition
Let us examine the practical workflow in greater detail, as institutions considering adoption will benefit from understanding the specific steps involved.
The hand-drawn sketch phase serves multiple functions. Sketching externalizes the designer's spatial intuitions in a form that can be analyzed and iterated upon. The sketch phase creates documentation of the human contribution that can be compared against later AI-generated material. Hand-drawing also establishes a baseline of intentionality that guides subsequent prompt engineering.
For organizations with existing drafting capabilities, the sketch phase integrates smoothly with current practice. The research does not require abandonment of analog methods but rather positions hand-drawing as essential to the hybrid workflow. Skilled draftsmanship retains value while gaining new collaborative possibilities.
The prompt engineering phase requires a particular kind of literacy that institutions may need to develop. Effective prompts that generate useful conceptual material demand understanding of how language translates into visual output through specific AI systems. The research describes prompt development as "iterative prompt engineering aligned with hand-drawn spatial visions." The iteration involves multiple rounds of prompt refinement, evaluation of outputs, and adjustment of language to guide the AI toward desired atmospheric and formal qualities.
Universities with design programs might consider developing dedicated coursework in prompt engineering for architectural applications. The skill set combines verbal precision, visual literacy, and understanding of algorithmic behavior. Prompt engineering represents a new kind of design competency that the field will increasingly require.
The AI-generated imagery then enters dialogue with physical modeling. The research produced translucent, petal-like forms that create meditative spatial experiences. Light behavior becomes a primary consideration, with the modeling process studying how illumination would move through and interact with proposed architectural elements.
The exhibition phase deserves particular attention from institutions considering the methodology. Public presentation serves as a validation mechanism, generating feedback that confirms or challenges the approach. The strong resonance at the Taiwan International Interior Design Expo suggests that audiences respond positively to work emerging from the hybrid human-AI process. For cultural institutions concerned about public reception of AI-assisted heritage work, the documented audience response provides useful reference.
Implications for Institutions: Cultural Ministries, Universities, and Design Organizations
The research carries distinct implications for different institutional categories, each worth examining in turn.
For government cultural ministries and heritage departments, Wu's methodology suggests new possibilities for sites that have resisted conventional preservation approaches. Abandoned sacred spaces often lack the historical significance required for major restoration funding while carrying deep meaning for local communities. A regeneration approach that honors memory while creating something new might find greater public support and more sustainable operational models than either full restoration or managed decline.
The hybrid workflow also offers possibilities for community engagement. The sketch phase could incorporate community input through participatory design sessions. The exhibition phase could serve as public consultation, gathering feedback before any permanent intervention proceeds. Governments seeking to democratize heritage decisions might find Wu's methodology particularly appealing.
For universities with architecture, design, or digital humanities programs, the research provides a documented case study for curriculum development. The combination of analog skill, AI literacy, theoretical framing, and public presentation creates a comprehensive pedagogical arc. Graduate programs seeking thesis project models or design studios developing coursework in emerging technologies could adopt and adapt Wu's approach.
The research also contributes to ongoing scholarly discourse about human-AI collaboration in creative fields. Academic institutions can engage with and build upon Wu's findings, extending the methodology to different site types, cultural contexts, or technological configurations.
For design studios and creative agencies, the research demonstrates how AI tools can enhance rather than replace human creativity. The positioning of AI as "generative partner" rather than automation tool offers a model for client communication. Studios can present AI-assisted work as augmentation of their creative capacity while maintaining clear human authorship and intentionality.
Those seeking to understand the complete methodology in detail can explore the full ai-human chapel regeneration research through the open-access publication available at ACDROI, which provides the complete peer-reviewed documentation including process images and exhibition outcomes.
The Pavilion Design: Spatializing Silence
The specific design outcome of Wu's research merits examination as the pavilion design demonstrates what the methodology can produce. The project, titled "Realm of Floating Silence," reimagines the abandoned chapel as a meditative pavilion where silence itself becomes spatialized.
The design employs layers of petal-like translucent forms that filter and modulate light. The translucent forms create what the research describes as an "immersive light behavior" where illumination becomes an active experiential element rather than mere visibility. Visitors moving through the space would encounter shifting qualities of luminosity as they progress.
The concept of "spatial rhythm" appears throughout the research, suggesting that movement through the pavilion follows a composed sequence rather than random exploration. The spatial rhythm connects to the original function of the chapel as a space of ritual, where procession and sequence carried meaning. The regenerated design maintains the quality of ritual progression while transforming the expression of that quality.
The translucent materiality serves multiple purposes. Translucent materials allow the surrounding environment to remain partially visible, maintaining connection between the renewed structure and the landscape context. The material choice creates interior atmospheres that shift with external conditions, ensuring the space remains dynamic rather than static. Translucent forms also produce a quality of lightness that contrasts with the typically massive construction of sacred architecture, suggesting transformation and ascension.
For institutions considering adaptation of the methodology, the specific formal outcome matters less than the process that generated the design. Different sites, different cultural contexts, and different program requirements would produce different designs. What transfers is the workflow, the analytical framework, and the collaborative positioning of AI tools.
The exhibition presentation at the Taiwan International Interior Design Expo included physical models alongside documentation of the design process. The combination of models and process documentation allowed audiences to understand the methodology while experiencing the outcomes. Institutions planning similar presentations might note the combined approach, which builds public understanding of AI collaboration while showcasing completed work.
Future Directions: AI Collaboration in Architectural Regeneration
Looking forward, Wu's research opens pathways that other researchers, institutions, and practitioners might pursue. The methodology demonstrated on a single chapel site could scale to different building types, different states of decay, and different cultural contexts.
Sacred spaces represent one category of abandoned architecture, but the hybrid workflow might apply equally to industrial heritage, institutional buildings, or residential structures with collective memory. Each category would require adaptation of the emotional, spatial, and ecological lenses to address different relationships between building and community.
The AI tools employed in Wu's research represent a particular moment in technological development. As generative AI capabilities evolve, the methodology will likely require updates to incorporate new possibilities. The underlying principles of treating AI as collaborative partner and maintaining human creative direction should remain stable even as specific tools change.
Institutional capacity building represents another future direction. Training programs that develop the specific competencies required for the hybrid workflow could expand the pool of practitioners capable of applying the methodology. Universities are well-positioned to develop training programs, potentially in partnership with cultural heritage organizations seeking to build internal capacity.
Documentation and knowledge sharing will prove essential. Wu's open-access publication of the research contributes to a growing body of knowledge about human-AI collaboration in design fields. As more practitioners adopt and adapt similar approaches, the cumulative documentation will provide increasingly robust guidance for institutions considering entry into the domain of AI-assisted heritage regeneration.
The research also suggests possibilities for participatory and community-centered extensions. The methodology demonstrated by a single designer could potentially be adapted for community co-design processes, with AI tools mediating between multiple human contributors rather than a single designer. The participatory extension would particularly interest institutions seeking to democratize heritage decision-making.
Closing Reflections
Wu's peer-reviewed research on poetic regeneration through AI-human collaboration offers institutions a documented framework for approaching abandoned sacred spaces with fresh creative possibilities. The hybrid workflow combining hand-drawn sketches, generative AI tools, physical modeling, and exhibition presentation demonstrates that technology can serve human creativity rather than supplanting human creative capacity. The emotional, spatial, and ecological lenses provide systematic structure for analyzing memory-laden sites while remaining open to the particular qualities of each location.
The strong public resonance documented at the Taiwan International Interior Design Expo suggests that audiences welcome the regeneration approach to heritage sites. For governments, universities, and design organizations seeking new models for cultural site stewardship, Wu's research provides valuable reference and inspiration.
As more sacred spaces worldwide fall silent, the question of how we might transform silence into renewed meaning grows increasingly urgent. What abandoned spaces in your institutional context might benefit from a methodology that treats decay, memory, and atmosphere as creative materials rather than obstacles to overcome?