Megan Waters
Group Leader, Research Library
Los Alamos National Laboratory
Finding a dataset is only the first step toward reusing it responsibly. Researchers and artificial intelligence (AI) systems also need to understand its provenance, limitations, access conditions, and permitted uses. This lightning talk examines a collaboration across US Department of Energy (DOE) laboratories to develop a shared data card: portable documentation combining structured metadata with narrative context for people and machines. Drawing on the technical report Data Cards for Standardized Metadata Across DOE-Aligned Data Initiatives, the presentation highlights three design lessons: establishing common metadata while accommodating local requirements; documenting sensitivity and use conditions separately for datasets and their metadata; and pairing automated validation with human stewardship. Released templates, a machine-readable schema, and validation tools provide a concrete foundation for this work. For library, repository, and research information technology leaders, the initiative offers an approach to connecting discovery, governance, and AI reuse across institutional boundaries. The talk invites consideration of the question of what context must travel with research data so that downstream users can judge whether and how it is reused.
https://www.osti.gov/biblio/3377514
https://www.energy.gov/undersecretaryforscience/genesis-mission/modcon-transformational-ai-and-data
https://www.energy.gov/undersecretaryforscience/genesis-mission/genesis-mission
