Navigating the AI-Driven Metadata Landscape: A Human Centered Approach

Part of the “Technology Matters” subtheme at the IFLA 2025 symposium, this presentation discusses the use of AI in metadata creation and management, based on findings from the Survey on Metadata and AI conducted by the DCMI Education Committee's Metadata and AI Task Group in late 2024. With 222 respondents—primarily from China and the U.S.—the survey explores the potential of generative and predictive AI tools, while also highlighting professional concerns. Respondents emphasized the importance of transparency, human oversight, and ethical considerations, particularly around bias in AI training data. While AI offers promising efficiency gains, the findings underline the need for continued professional development and robust evaluation frameworks. Controlled vocabularies, community review, and clearly defined workflows are essential to ensure trust in AI-assisted metadata. Ongoing data collection may yield further insights for future discussion, contributing to the development of best practices for integrating AI and human contributions in metadata work.