Metadata for the Margins: Cyberpunk Cataloging with OpenAI

Abstract

Standardized taxonomies and controlled vocabularies form the foundation of universal bibliographic control. Through our practice-based case study, we demonstrate how artificial intelligence bridges the gap between global standards and local needs for specialized and complementary collections that do not fit neatly into conventional cataloging taxonomies. We present a real-world application of an open access, efficient, cost effective, and context-sensitive AI-driven solution for cataloging non-bibliographic resources, with a focus on customizable templates tailored to specific item types.

The templates we present are seamlessly integrated with AI to import descriptive metadata including images and videos in one record. In addition to streamlining metadata creation, the system supports multilingual input and output, enabling the generation of accurate metadata in different languages for culturally specific or regionally unique collections. We will show examples to highlight the flexibility and adaptability of the system to save data in different formats, emphasizing its capacity to support locally contextualized subject headings and to streamline workflows. The use of web-based AI populated templates also simplify staff training by enabling non-experts to create structured, reliable cataloging records fostering capacity-building in institutions lacking traditional cataloging expertise. Finally, we will discuss the potential applications and technical requirements necessary for implementation, ensuring that libraries across diverse institutional landscapes can adopt this solution, regardless of staffing or funding constraints. This presentation is ideal for librarians seeking innovative, cost-effective tools to enhance the discoverability of their diverse collections.

We contribute a replicable, AI-driven, and human-centered metadata tool that promotes access, equity, and sustainability for every collection, no matter how small, unconventional, or underfunded has a place in the global metadata network, thereby strengthening the foundation of universal bibliographic control through reliable data to make libraries stronger.

(presented on 15 August 2025 at "Pushing Boundaries to Next Generation Cataloguing: Experiments at the Edge of AI and Metadata" session)