University of Michigan (University of Michigan, School of Information)

The University of Michigan School of Information and the Inter-university Consortium for Political and Social Research (ICPSR) will study how generative artificial intelligence (GenAI) might enhance data curation. A prior IMLS-funded study revealed that nationally, the data curation field is challenged with the time-intensive nature of generating expert-crafted, detailed metadata and addressing potential disclosure risks when sharing data for reuse. This research project will pursue both how GenAI can aid in drafting and evaluating metadata, as well as how computational tools can assist in identifying and managing sensitive data elements. Evaluation of GenAI tools’ effectiveness will involve curator feedback, user studies, text analytics, and comparisons with ICPSR’s current metadata curation and search indices. Project beneficiaries across the nation will include data curators gaining efficiency in their roles; data reusers enjoying better data access and variable-level, detailed metadata; and research participants benefiting from enhanced data privacy through reduced disclosure risks.