Framework for Camera-Assisted Book Classification Using ChatGPT and Library of Congress Classification System in libraries

This study proposes a camera-assisted framework for book classification in academic libraries using ChatGPT and the Library of Congress Classification (LCC) system, with specific relevance to Nigerian university libraries. Manual assignment of LCC call numbers, though accurate, is time-consuming and contributes to cataloguing backlogs due to staff shortages and growing collections. Drawing on advances in artificial intelligence, optical character recognition (OCR), and computer vision, the study explores how multimodal technologies can support professional cataloguing workflows without undermining established standards. A developmental research design was adopted, comprising needs assessment, framework design, prototype development, and evaluation. The proposed framework integrates smartphone-based image capture of book covers and title pages, OCR and computer vision for text extraction, structured prompt engineering with ChatGPT, heuristic mapping to LCC schedules, and mandatory human verification. A functional prototype was implemented using accessible tools suitable for bandwidth-constrained environments and evaluated through feasibility, accuracy, and usability testing with professional cataloguers. Findings indicate that the framework can significantly reduce initial classification time and support consistent subject analysis, while maintaining compliance with LCC through librarian oversight. AI-generated suggestions were generally accurate at class and subclass levels but required human validation for interdisciplinary and context-specific materials. The study demonstrates that camera-assisted AI tools can augment, rather than replace, cataloguers by automating routine tasks and improving workflow efficiency. The study contributes a practical, context-sensitive model for responsible AI integration into technical services in Nigerian academic libraries and provides a foundation for future empirical studies on multimodal classification systems.