Metadata
Working Group
The Metadata Working Group will explore and share the ways in which AI technologies can be applied to the creation and enhancement of metadata for cultural heritage resources. It will also address the role of metadata in collecting, curating, and describing the resources needed for AI, machine learning, and data science more broadly.

AI technologies are transforming how metadata in libraries, archives, and museums is created, enriched, and critically evaluated, enhancing both access to collections and the understanding of AI systems themselves.
Over the years, libraries, archives, and museums have invested significant resources into creating metadata to describe and manage their collections. AI technologies may provide opportunities to enrich the collective metadata of LAM institutions and foster enhanced functionality, including discovery of these institutions’ collections and services. Generative AI facilitates the creation and generation of new metadata, making workflows more efficient and enabling novel forms of access to content. Conversely, metadata professionals can bring their expertise in resource description to bear on the AI domain itself, describing and documenting AI and ML models and datasets and identifying areas of bias that could harm already marginalized groups and evaluating the output of AI.
Monthly Call – The meetings takes place on the second Tuesday of each month with a duration of one hour.
15:00 UTC

Information and demonstration of AI in Metadata workflows, research projects, and general hub for people interested in the practice and application of AI related to metadata.
- Reports, white papers, or literature reviews.
- Development and testing of new tools and workflows.
- Presentations by GLAM and Computer Science Professionals
- Schemas or ontologies for describing and documenting ML models and datasets.
- Use of AI in the FOLIO Library Services Platform
- Use of AI with generation and analysis of BIBFRAME Linked Data
- Adding AI functionality to the Blue Core shared bibliographic datastore
- Use of AI in digital libraries for different media types
- Non-MARC machine learning tasks for finding aids

