From ChatGPT to CatGPT: The Implications of Artificial Intelligence on Library Cataloging
Open accessAbstract
This paper explores the potential of language models such as ChatGPT to transform library cataloging. Through experiments with ChatGPT, the author demonstrates its ability to generate accurate MARC records using RDA and other standards such as the Dublin Core Metadata Element Set. These results demonstrate the potential of ChatGPT as a tool for streamlining the record creation process and improving efficiency in library settings. The use of AI-generated records, however, also raises important questions related to intellectual property rights and bias. The paper reviews recent studies on AI in libraries and concludes that further research and development of this innovative technology is necessary to ensure its responsible implementation in the field of library cataloging.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The accuracy failures documented by Chow et al. (2024), Brzustowicz (2023), and Tang & Jiang (2025) are not primarily failures of topic identification; they are failures of rule application.
“The accuracy failures documented by Chow et al. (2024), Brzustowicz (2023), and Tang & Jiang (2025) are not primarily failures of topic identification-they are failures of rule application.”
From A Skill-Based AI Agentic Pipeline for Library of Congress Subject Indexing (2026) · cites this paper as Brzustowicz (2023)
Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
Experimental — accuracy not yet measuredUnlike single-prompt LLM approaches, the system explicitly encodes SHM rules at each stage rather than relying on the model's internalized knowledge of LCSH conventions.
“Unlike the single-prompt LLMapproaches of Brzustowicz (2023) and Chow et al. (2024), it explicitly encodes SHM rules at each stage rather than relying on the model's internalized knowledge of LCSH conventions.”
From A Skill-Based AI Agentic Pipeline for Library of Congress Subject Indexing (2026) · cites this paper as Brzustowicz (2023)
The Methodology (pp. 2–3) describes asking ChatGPT to generate catalog records with individual requests, supporting the characterization of Brzustowicz’s experiments as single-prompt approaches. It also quotes ChatGPT describing training on MARC and RDA rules, but does not establish that the approach specifically relied on internalized LCSH conventions rather than explicit SHM rules. (LLM, conf=0.82)
Experimental — accuracy not yet measuredNotable discrepancies emerged in the assignment of subject access points by ChatGPT.
“However, notable discrepancies emerged in the assignment of subject access points, suggesting that while LLMs internalize general cataloging patterns from their training data, they lack the systematic application of LCSH policy rules required for reliable subject heading construction.”
From A Skill-Based AI Agentic Pipeline for Library of Congress Subject Indexing (2026) · cites this paper as Brzustowicz (2023)
ChatGPT could produce records comparable to those created by professional catalogers for basic metadata fields (title, author, publisher).
“The results showed that ChatGPT could produce records comparable to those created by professional catalogers for basic metadata fields (title, author, publisher), and could even generate original records for items without existing WorldCat entries.”
From A Skill-Based AI Agentic Pipeline for Library of Congress Subject Indexing (2026) · cites this paper as Brzustowicz (2023)
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Chatting about ChatGPT: how may AI and GPT impact academia and libraries?
Lund, B. D., & Wang, T. (2023) · Library Hi Tech News · doi:10.1108/lhtn-01-2023-0009 · Cited by 1,355 · see citing works
ChatGPT is not all you need. A state-of-the-art review of large generative AI models
Gozalo-Brizuela, R., & Garrido-Merchan, E. C. (2023) · arXiv · doi:10.48550/arxiv.2301.04655 · Cited by 245 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Portillo, I., & Carson, D. (2025) · Journal of the Medical Library Association · doi:10.5195/jmla.2025.2079 · Cited by 5 · see citing works
Peekhaus, W. (2026) · The Journal of Academic Librarianship · doi:10.1016/j.acalib.2025.103198 · Cited by 3 · see citing works
A Skill-Based AI Agentic Pipeline for Library of Congress Subject Indexing
Chow, E. H. C. (2026) · arXiv · doi:10.48550/arxiv.2605.03537
Bibliographic Metadata Extraction Using Deterministic Retrieval and Specialized AI Agents
Rasheva-Yordanova, K., Hadzhiyska, Y., Ivanov, I., Kostadinova, I., Lozev, L., & Lozeva, A. (2026) · Information · doi:10.3390/info17090841
- Authors
- Brzustowicz, R.
- Affiliations
- Carlow University (US)
- Corresponding author
- Brzustowicz, Richard
- ORCID
- Brzustowicz, Richard
- Venue
- Information Technology and Libraries
- Publisher
- American Library Association
- Published
- 2023-09-18
- Volume
- 42
- Issue
- 3
- ISSN
- 0730-9295 (ISSN-L)
- 2163-5226
- OpenAlex
- W4387313225
- Cited by
- 57
- References
- 2
- Citations by year
- 2026: 24 · 2025: 22 · 2024: 11
- FWCI
- 13.38
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- Locations
- Information Technology and Libraries (publishedVersion, cc-by-nc)
- Information Technology and Libraries (Boston College) (submittedVersion, cc-by-nc)
- Topics
- Physical Sciences › Computer Science › Information Systems › Library Science and Information Systems
- Physical Sciences › Computer Science › Information Systems › Library Science and Information
- Physical Sciences › Computer Science › Information Systems › Web and Library Services
- Keywords
- ChatGPT, library cataloging, MARC records, Resource Description and Access, Dublin Core, large language models, intellectual property, algorithmic bias, responsible AI use
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
