Subject Indexing in Libraries: Shifting Responsibilities Between Humans and Algorithms

In the context of limited computing power and small training sample sizes, the challenge of achieving high-quality subject indexing is particularly acute. The human-in-the-loop (HITL) concept offers a solution that combines automated processing with expert human input. The novelty of the proposed approach lies in the use of active learning and targeted tagging mechanisms, which enables indexing accuracy comparable to that of full-text neural network models with minimal resource expenditure. An analysis of the distribution of functions between process participants demonstrates that, in the era of algorithmic methods, human intelligence and library classification retain their role as guarantors of information quality and reliability.

Within this model, the librarian transitions from manual indexing to quality control and system training, verifying algorithm results and correcting errors. A subject expert generates training samples and provides interpretation of the scientific context for the system. The programmer, in turn, develops an interface for making edits and implements a closed feedback loop that allows for further model training.

A detected index error is interpreted not as an individual cataloger’s oversight, but as a signal to recalibrate the interactions of all participants in the process. It has been established that manually structured datasets outperform automated counterparts in terms of the effectiveness of training neural network models. Libraries’ efforts to create and maintain high-quality metadata not only remain relevant but are also becoming a key factor in ensuring the quality of artificial intelligence systems. In today’s environment, a library’s value is determined not by the size of its archival collections, but by its ability to transform information into high-quality datasets suitable for solving modern scientific problems.

1. Zakharova S.S. Possibilities of Artificial Intelligence in Compiling a Subject-Thematic Index, Bibliosfera [Bibliosphere], 2025, no. 4, pp. 82—88. DOI: 10.20913/1815-3186-2025-4-82-88 (in Russ.).

2. Kasprzik A. The Automation of Subject Indexing at ZBW and the Role of Metadata in Times of Large Language Models, Procedia Computer Science, 2024. vol. 249, pp. 160—166. DOI: 10.1016/j.procs.2024.11.059.

3. Kasprzik A. Transferring Applied Machine Learning Research into Subject Indexing Practice, New Horizons in Artificial Intelligence in Libraries. Berlin, De Gruyter Saur Publ., 2025, pp. 199—212. DOI: 10.1515/9783111336435-015.

4. Han K., Rezapour R., Nakamura K., Devkota D., Miller D.S., Diesner J. An Expert-in-the-Loop Method for Domain-Specific Document Categorization Based on Small Training Data, Journal of the Association for Information Science and Technology, 2023, vol. 74, no. 6, pp. 669—684. DOI: 10.1002/asi.24714.

5. Holzinger A., Plass M., Holzinger K., Crişan G., Pintea C.-M., Palade V. A Glass-Box Interactive Machine Learning Approach for Solving NP-Hard Problems with the Human-in-the-Loop, Creative Mathematics and Informatics, 2019, vol. 28, no. 2, pp. 121—134. DOI: 10.37193/CMI.2019.02.04.

6. Monarch (Munro) R. Human-in-the-Loop Machine Learning: Active Learning and Annotation for Human-Centered AI. Moscow, DMK Press Publ., 2022, 497 p. (in Russ.).

7. Poley C., Uhlmann S., Busse F., Jacobs J.-H., Kähler M., Nagelschmidt M., Schumacher M. Automatic Subject Cataloguing at the German National Library, Liber Quarterly, 2025, vol. 35, no. 1, pp. 1—29. DOI: 10.53377/lq.19422.

8. Kunze H. Sostavlenie vspomogatel’nykh ukazatelei [Compilation of Auxiliary Indexes]. Moscow, Kniga Publ., 1977, 62 p.

9. Prizment E.L., Dinershtein E.A. Vspomogatel’nye ukazateli k knizhnym izdaniyam [Auxiliary Indexes to Book Publications]. Moscow, Kniga Publ., 1988, 207 p.

10. Ahmed M., Mukhopadhyay M., Mukhopadhyay P. Automated Knowledge Organization: AI/ML-Based Subject Indexing System for Libraries, DESIDOC Journal of Library & Information Technology, 2023, vol. 43, no. 1, pp. 45—54. DOI: 10.14429/djlit.43.01.18619.

11. Hanegan M., Rosser C. Generative AI and Libraries: Claiming Our Place in the Center of a Shared Future. Chicago, ALA Editions Publ., 2025, 160 p.

12. Sokolova J.V. Semantic Treatment of Documents on Soil Science, Kul’tura: teoriya i praktika: ehlektronnyi nauchnyi zhurnal [Culture: Theory and Practice], 2025, no. 4 (65). Available at: http://theoryofculture.ru/issues/142/1720 (accessed 20.04.2026) (in Russ.).

13. Melnikova E.V., Tsvetkova V.A. Analysis of Possibilities of Artificial Intelligence Application in Modern Scientometry and Bibliometry, Vestnik RGGU. Seriya: Informatika. Informatsionnaya bezopasnost’. Matematika [RGGU Bulletin. Series: Information Science. Information Security. Mathematics], 2025, no. 2, pp. 19—40. DOI: 10.28995/2686-679X-2025-2-19-40 (in Russ.).

14. Sakovich D.A. The Potential of Artificial Intelligence in Document Cataloging, Biblioteki v informatsionnom obshchestve: sokhranenie traditsii i razvitie novykh tekhnologii: doklady VI Mezhdunarodnoi nauchnoi konferentsii, Minsk, 5—6 dekabrya 2024 g. [Libraries in the Information Society: Preserving Traditions and Developing New Technologies: Proceedings of the 6th International Scientific Conference Minsk, December 5—6, 2024]. Minsk, 2024, pp. 84—91. DOI: 10.5281/zenodo.14162068 (in Russ.).

15. Pistsov S.M., Trusov V.A. Conceptual Models of Semantic Processing of Scientific and Technical Information, Nauchno-tekhnicheskaya informatsiya. Seriya 2: Informatsionnye protsessy i sistemy [Automatic Documentation and Mathematical Linguistics], 2025, no. 3, pp. 19—27. DOI: 10.36535/0548-0027-2025-03-3 (in Russ.).

16. Mosqueira-Rey E., Hernández-Pereira E., Alonso-Ríos D., Bobes-Bascarán J., Fernández-Leal A. Human-in-the-Loop Machine Learning: A State of the Art, Artificial Intelligence Review, 2023, vol. 56, pp. 3005—3054. DOI: 10.1007/s10462-022-10246-w.

17. Moiseeva N.A. Artificial Intelligence Technologies in Information and Library Systems, Nauchnye i tekhnicheskie biblioteki [Scientific and Technical Libraries], 2024, no. 5, pp. 85—101. DOI: 10.33186/1027-3689-2024-5-85-101 (in Russ.).

18. Romero-Obon M., Rouaz-El-Hajoui K., Sancho-Ochoa V., Vargas R., Pérez-Lozano P., Suñé-Pou M., García-Montoya E. Human-in-the-Loop AI Use in Ongoing Process Verification in the Pharmaceutical Industry, Information, 2025, vol. 16, no. 12, art. 1082, 19 p. DOI: 10.3390/info16121082.

19. Noruzi A. The Use of Artificial Intelligence in Knowledge Organization and Subject Indexing, Informology, 2024, vol. 3, issue 1, pp. 1—8.

20. Mitroshin I.A. Avenues for Artificial Intelligence in Library and Information Services, Nauchnye i tekhnicheskie biblioteki [Scientific and Technical Libraries], 2025, no. 1, pp. 120—134. DOI: 10.33186/1027-3689-2025-1-120-134 (in Russ.).

21. Stepanov V.K. Artificial Intelligence in Libraries: Theoretical Approaches and Practical Solutions (on the Conclusions of the Scientific and Practical Conference “Artificial Intelligence in Library and Information Services”), Nauchnye i tekhnicheskie biblioteki [Scientific and Technical Libraries], 2024, no. 11, pp. 15—30. DOI: 10.33186/1027-3689-2024-11-15-30 (in Russ.).

22. Goncharov M.V., Sokolinsky K.E., Shraiberg Ya.L. Artificial Intelligence in the Practice of Sci-Tech Libraries: The Study of Potential, Experience and Prospects Evaluation, Nauchnye i tekhnicheskie biblioteki [Scientific and Technical Libraries], 2025, no. 12, pp. 144—164. DOI: 10.33186/1027-3689-2025-12-144-164 (in Russ.).

23. Stepanov V.K. Artificial Intelligence in Library and Information Activities: One Year Later, Nauchnye i tekhnicheskie biblioteki [Scientific and Technical Libraries], 2025, no. 12, pp. 127—143. DOI: 10.33186/1027-3689-2025-12-127-143 (in Russ.).