Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 33 of the 34 references it cites were found in a registry or an index; 29 of the 64 sources its claims cite were checked for support.

Fabrication and errors in the bibliographic citations generated by ChatGPT

Open access
Walters, William H., Wilder, Esther Isabelle · 2023 · Scientific Reports

Abstract

Although chatbots such as ChatGPT can facilitate cost-effective text generation and editing, factually incorrect responses (hallucinations) limit their utility. This study evaluates one particular type of hallucination: fabricated bibliographic citations that do not represent actual scholarly works. We used ChatGPT-3.5 and ChatGPT-4 to produce short literature reviews on 42 multidisciplinary topics, compiling data on the 636 bibliographic citations (references) found in the 84 papers. We then searched multiple databases and websites to determine the prevalence of fabricated citations, to identify errors in the citations to non-fabricated papers, and to evaluate adherence to APA citation format. Within this set of documents, 55% of the GPT-3.5 citations but just 18% of the GPT-4 citations are fabricated. Likewise, 43% of the real (non-fabricated) GPT-3.5 citations but just 24% of the real GPT-4 citations include substantive citation errors. Although GPT-4 is a major improvement over GPT-3.5, problems remain.

Overview

CoChat Summary

## TL;DR

This study tested how often ChatGPT fabricates bibliographic citations and how often its citations to real works contain errors. Across 636 citations in 84 short papers, GPT-4 improved over GPT-3.5 but still produced substantial fabrication and citation-error rates.

## Key findings

  • In 84 ChatGPT-generated papers on 42 topics, the 636 citations were split between GPT-3.5 and GPT-4 outputs; 93% of cited works were scholarly rather than popular.
  • Of the 222 works cited in the GPT-3.5 papers, 55% were fabricated; for GPT-4, only 18% were fabricated.
  • Among real citations, 43% of GPT-3.5 citations had one or more substantive citation errors, compared with 24% for GPT-4.
  • More than 40% of citations had minor APA formatting errors, and fewer than 10% of citations to real works included hyperlinks; when hyperlinks were present in real citations, about one-third were inaccurate.

## Methods

  • GPT-3.5 and GPT-4 each generated 42 short literature-review papers from prompts on multidisciplinary topics, yielding 84 papers total and 636 cited works.
  • The authors searched Google, Google Scholar, Amazon, DOAJ, PubMed, Scopus, WorldCat, publisher/journal sites, and other sources to determine whether each citation was real or fabricated.
  • For real works, they checked substantive citation errors, APA-format adherence, and whether hyperlinks were present and led to the cited work.

## Limitations

  • The paper states that it cannot know what proportion of users are taking advantage of GPT-4’s improved performance because detailed information on ChatGPT use is not available.
  • The papers were short and generated in response to a specific prompt style, so the results may not generalize beyond this setup; this is evident from the text.
  • The study notes that any text after the initial bibliography was excluded, which may omit additional citation behavior; this is evident from the methods.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41598-023-41032-5

Cites

33 of 34 references verified1 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

11 Valid · 0 Not valid · 11 Not checked · 49 Own results and interpretations

38 more claims · 2 with their cited sources checked

Create a free account to see all 71 claims and to verify the 3 references still open.

See all claims — free account
  • Probably not a substantial claim. 17
  • It may state two separate claims as one. 13

Create a free account to see these claims.

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?

33 more claims about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

Citations

Cites

Show all 31 references

Research works printed but not linked: 3

  1. DePeau-Wilson, M. Peer-reviewed journal publishes paper written almost entirely by ChatGPT. MedPage Today. https://www.medpagetoday.com/special-reports/exclusives/102960 (2023).
  2. Atlas, S. Chatbot Prompting: A Guide for Students, Educators, and an AI-Augmented Workforce (S. Atlas, 2023). https://www.researchgate.net/publication/367464129_Chatbot_Prompting_A_guide_for_students_educators_and_an_AI-augmented_workforce.
  3. Jacsó, P. As we may search: Comparison of major features of the Web of Science, Scopus, and Google Scholar citation-based and citation-enhanced databases. Curr. Sci. 89, 1537–1547 (2005).

Cited by

See all 499

Metadata

Authors
Walters, W. H., & Wilder, E. I.
Affiliations
  • Manhattan University (US)
  • The Graduate Center, CUNY (US)
Corresponding author
Walters, William H.
Venue
Scientific Reports
Published
2023-09-07
Volume
13
Issue
1
Pages
14045-14045
ISSN
2045-2322 (ISSN-L)
OpenAlex
W4386510404
Cited by
499
References
31
Citations by year
2026: 249 · 2025: 117 · 2024: 118 · 2023: 12
FWCI
123.75
Type
journal article
Language
en
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • CUNY Academic Works (City University of New York) (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Social Sciences › Education › Academic integrity and plagiarism
  • Social Sciences › Library and Information Sciences › Academic Writing and Publishing
  • Social Sciences › Library and Information Sciences › scientometrics and bibliometrics research
Keywords
ChatGPT, large language models, hallucinations, bibliographic references, citation errors, GPT-3.5, GPT-4

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.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-09-28, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-09-28 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Fabrication and errors in the bibliographic citations generated by ChatGPT. https://app.cochat.ai/papers/10.1038/s41598-023-41032-5
The paper
Walters, W. H., & Wilder, E. I. (2023). Fabrication and errors in the bibliographic citations generated by ChatGPT. Scientific Reports. https://doi.org/10.1038/s41598-023-41032-5

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.