Measuring the Effectiveness of Artificial Intelligence Applications in Book Classification

A Comparative Study with the Library of Congress Book Classification in Terms of Accuracy and Performance

Abstract

With the increasing use of artificial intelligence (AI) technologies in various fields, important questions have arisen regarding their ability to perform precise tasks that require human expertise, such as classifying books in libraries. This study aims to assess the accuracy of AI in this task and compare it with traditional classification methods used by the Library of Congress, which serves as a global standard for precision.  Through analysis, it was found that the AI model [ChatGPT] achieved satisfactory results in classifying books within the humanities and social sciences, with classification numbers closely aligning with those used by the Library of Congress. However, the model faced notable difficulties in fields such as medicine and technology, which require precise identification of subfields and a deep understanding of specialized terminology.  In light of these findings, the study recommends against relying entirely on AI for classification tasks. Instead, AI should be used as an assistant to expedite routine or large-scale tasks, particularly in cases requiring fast processing. The study emphasizes that human experts should retain the primary role in reviewing results and ensuring their accuracy, as human intelligence remains superior in understanding nuanced details and complex contexts. By combining the efficiency and speed of AI with human precision and expertise, an optimal balance can be achieved to enhance the quality of work in libraries without compromising absolute accuracy.

Through analysis, it was found that the AI model [ChatGPT] achieved satisfactory results in classifying books within the humanities and social sciences, with classification numbers closely aligning with those used by the Library of Congress. However, the model faced notable difficulties in fields such as medicine and technology, which require precise identification of subfields and a deep understanding of specialized terminology. 

In light of these findings, the study recommends against relying entirely on AI for classification tasks. Instead, AI should be used as an assistant to expedite routine or large-scale tasks, particularly in cases requiring fast processing. The study emphasizes that human experts should retain the primary role in reviewing results and ensuring their accuracy, as human intelligence remains superior in understanding nuanced details and complex contexts. By combining the efficiency and speed of AI with human precision and expertise, an optimal balance can be achieved to enhance the quality of work in libraries without compromising absolute accuracy.