LLM-Based Creativity and Computing in the Preservation of Cultural Heritage

Dear Colleagues,

Cultural heritage preservation is entering a new computational era in which large language models (LLMs), multimodal generative systems, knowledge graphs, and human-centred creative computing can work together to document, interpret, restore, and communicate cultural memory. Tangible and intangible heritage, including historic manuscripts, archaeological records, museum collections, oral histories, endangered languages, rituals, music, architecture, and craft traditions, often exists in fragmented, multilingual, and multimodal forms. Conventional digitisation has made major progress in archiving these materials, but preservation now requires richer methods for semantic understanding, contextual reconstruction, responsible interpretation, and participatory access.

LLM-based creativity and computing provide new opportunities for this field. LLMs can support the transcription and translation of historical texts, generate metadata for heterogeneous archives, assist curators in building interpretive narratives, and connect dispersed evidence across collections. When combined with computer vision, 3D reconstruction, augmented and virtual reality, digital twins, and interactive storytelling, these models can help produce immersive and educational experiences that make cultural heritage more accessible to scholars, communities, and the public. At the same time, heritage contexts demand careful attention to authenticity, provenance, cultural sensitivity, copyright, indigenous data sovereignty, bias, and the risk of hallucinated or over-simplified narratives.

This Topic will spotlight research that uses LLMs and creative computing to advance the preservation, analysis, restoration, and communication of cultural heritage. Key themes include:

• Computational preservation and documentation: LLM-assisted annotation, cataloguing, transcription, translation, summarisation, and semantic linking of cultural heritage materials.
• Creative interpretation and public engagement: generative storytelling, interactive museum experiences, educational agents, and immersive cultural heritage applications.
• Responsible and trustworthy heritage AI: provenance-aware generation, expert-in-the-loop validation, bias mitigation, explainability, cultural ethics, and community-centred governance.

By bringing together artificial intelligence, creative computing, digital humanities, museum studies, archaeology, information science, and heritage management, this Topic aims to promote rigorous and culturally responsible computational methods that preserve the past while enabling new forms of creative access and interpretation.

Prof. Dr. Kelvin K. L. Wong
Prof. Dr. Adrian David Cheok
Prof. Dr. Dhanjoo N. Ghista
Topic Editors