From Reconstruction to Reimagination
Creative Uses of AI-Generated 3D
AI and 3D imaging are often framed as tools for reconstruction, to recreate what was lost, document what still exists, or preserve what might soon disappear. But increasingly, they’re also being used for something more ambitious: reimagination.
Across cultural heritage, digital arts, and education, generative 3D tools are not just restoring the past, they’re helping us explore alternate versions of it, reinterpret meaning, and make the invisible visible. The result is a shift from documentation to dialogue, where technology doesn’t just record history, but opens new ways of engaging with it.
At Aralia, we’re seeing this potential emerge through projects that blend Elata’s AI-powered 3D imaging with speculative, educational, and creative applications.
Reconstruction vs Reimagination
Let’s be clear, faithful 3D reconstruction is still vital. From heritage preservation to disaster recovery, the ability to capture and recreate lost structures or artefacts with high fidelity remains an essential use case.
But reimagination asks a different question:
“What if we could visualise not just what was, but what might have been?”
It’s a space where historians collaborate with designers, where archaeologists work with artists, and where communities contribute their own interpretations. And it’s a space that generative AI and XR are uniquely positioned to support.
Speculative Heritage and Creative Interpretation
With AI-assisted modelling, we can now generate visual hypotheses about:
How a ruined structure may have looked before collapse
How artefacts relate to others in distant collections
How alternative architectural choices could have shaped a space
In one recent conversation, a curator described using generative 3D to visualise speculative roof structures for a Roman villa, allowing them to compare reconstructions based on different interpretations of the evidence. The result wasn’t just a “finished model” but a tool for discussion, engagement, and critical thinking.
AI doesn’t give us answers, it gives us starting points for richer questions.
Engaging the Public with New Perspectives
By blending real scans with AI-generated extrapolations, cultural organisations can offer new ways to engage audiences:
Interactive galleries that explore different reconstruction theories
XR experiences where users can toggle between versions of a site over time
Community cocreation platforms that invite local people to “fill in the gaps” based on oral history or family memory
This turns static displays into participatory, evolving narratives, where heritage isn’t fixed, but layered with meaning, memory, and imagination.
Responsible Creativity with Generative AI
Of course, creative use of generative AI must be clearly framed and ethically managed. Visitors and users should always understand what’s real, what’s extrapolated, and what’s speculative.
We advocate for:
Transparent labelling of AI-generated content
Versioning that preserves original scans alongside imaginative outputs
Collaboration with domain experts to ground speculative models in evidence
Used responsibly, generative 3D doesn’t replace the real, it adds new dimensions to how we think about it.
Final Thoughts: The Past Isn’t Finished
As tools like Elata make 3D imaging more accessible, we’re excited by the shift from preservation to participation, from reconstructing the past to reimagining it together.
Creative uses of AI-generated 3D won’t replace traditional heritage methods. But they will enrich them, making space for new stories, new interpretations, and new audiences.
Because heritage isn’t just what we inherit, it’s what we continue to shape.
