GLAMorous AI TL;DR — March 2026

This edition explores how AI is being applied across heritage science, conservation, and collections research, with a focus on data integration, metadata, and reuse.

🔬 AI for Heritage Science Data Integration

Li et al. — Machine Learning and Deep Learning for Cultural Heritage Conservation: A Bibliometric and Task-Oriented Review (2026)
A major review showing how AI is used across recognition, reconstruction, and monitoring, but emphasising that success depends on how well data and tasks are aligned.
Read more →

Baek et al. — Development of an Artificial Intelligence-based Platform for the Analysis and Utilization of Cultural Heritage Data (2025)
Introduces an AI platform combining archives, analytics, and generative tools, including relational models that link complex heritage datasets together.
Read more →

What this means:
AI can connect heritage science datasets across instruments and projects, but only when data structures and metadata are consistent. Otherwise, it simply scales fragmentation.

🧠 AI + Metadata, Ontologies & Knowledge Graphs

Felicetti et al. — Knowledge Graphs and Artificial Intelligence for the Implementation of Cognitive Heritage Digital Twins (2025)
Demonstrates how AI can transform fragmented heritage documentation into structured, interoperable knowledge graphs aligned with CIDOC CRM.
Read more →

Ignatowicz et al. — Position Paper: Metadata Enrichment Model: Integrating Neural Networks and Semantic Knowledge Graphs for Cultural Heritage Applications (2025)
Combines computer vision, LLMs, and semantic graphs to generate richer metadata for digitised collections, highlighting the need for domain-specific training.
Read more →

📍 Why this matters:
AI can generate and link metadata, but interoperability depends on ontologies and standards, not automation alone.

🧪 AI for Materials Analysis & Conservation

Pouyet et al. — Artificial Intelligence for Pigment Classification Task in the Short-Wave Infrared Range (2021)
Machine learning models can classify pigments and mixtures from spectral data, supporting non-invasive analysis of artworks.
Read more →

Colace et al. — New AI challenges for cultural heritage protection: A general overview
A review of how AI is being applied to detect damage, monitor sites, and support restoration using imaging and sensor data. While technically effective, most approaches remain isolated from broader data infrastructures and lack integration with metadata and governance frameworks.
Read more →

Cao — The Transformation of Traditional Chinese Painting in the Digital Art Wave: The Impact of AIGC and Virtual Reality (2026)
This review examines how generative AI and virtual reality are reshaping traditional Chinese painting through style generation, immersive viewing, and digital reconstruction. It argues that while these tools can reproduce visual features convincingly, they still struggle to capture deeper cultural meaning, artistic intention, and contemplative ways of seeing.

Read more →

📍 Critical note:
These approaches work best in controlled datasets. Real heritage materials introduce noise, degradation, and uncertainty that models struggle to generalise.

🔍 AI for Legacy Data & Scientific Documentation

Felicetti et al. — AI Workflows for Extracting Structured Knowledge (2025)
Shows how AI can extract structured data from unstructured heritage science documentation and integrate it into knowledge graphs.
Read more →

Casillo et al. — AI in Archaeological and Heritage Conservation (2025)
Reviews AI applications in monitoring, site analysis, and conservation, emphasising integration challenges across datasets.
Read more →

📍 Key challenge:
AI can extract entities and relationships, but it cannot reconstruct missing context or undocumented methodology.

❓ Big Question

If AI can:

• classify materials

• link datasets

• generate metadata

• extract knowledge from reports

But cannot reliably:

• interpret context

• resolve ambiguity

• enforce standards

Is the real bottleneck in heritage science not analysis, but data governance?

💬 About

I’m Alfie, a heritage scientist working at the intersection of AI, data governance, and heritage science infrastructure (REVEAL / RICHeS).

This digest focuses on what AI actually does in practice in heritage science, not what it promises.

👉 Read or subscribe at glamorousai.beehiiv.com
👉 Send papers, tools, or reports for the next edition