The DOI matches its Crossref record; 37 of the 40 references it cites were found in a registry or an index; 19 of the 43 sources its claims cite were checked for support.
Fabricated or accurate? Ethical concerns and citation hallucination in aI-generated scientific writing on musculoskeletal topics
Open accessAbstract
Aims: Large language models (LLMs) such as ChatGPT are increasingly used in academic and clinical writing. While these tools can generate coherent and domain-specific text, concerns persist regarding the accuracy of their automatically generated references. In musculoskeletal rehabilitation—a field heavily reliant on current evidence—the reliability of citations is especially critical. Yet, systematic evaluations of citation accuracy in AI-generated scientific content are lacking. To evaluate the reference accuracy of scientific texts generated by ChatGPT (GPT-4) in response to musculoskeletal rehabilitation prompts, and to determine whether reference accuracy improves following structured post-generation verification. Methods: ChatGPT was prompted to generate four scientific paragraphs on musculoskeletal rehabilitation topics (manual therapy, ACL reconstruction, low back pain, and rotator cuff repair), each including 10 references with DOIs. A total of 40 references were analyzed using a 3-point scoring system (0=fabricated, 1=partially correct, 2=fully accurate), which was used to assess citation quality. After initial evaluation, ChatGPT was asked to verify and revise its references. Scores before and after this step were compared descriptively and with Wilcoxon signed-rank tests to assess statistical significance, and effect sizes (r) were calculated to estimate the magnitude of improvement. Results: Only 7.5% of references were fully accurate in the initial generation, while 42.5% were completely fabricated. The remaining 50% were partially correct. After verification, the proportion of fully accurate references rose to 77.5%. Wilcoxon signed-rank testing confirmed a statistically significant improvement in accuracy across all prompts (W=561.0, p
Overview
CoChat Summary
TL;DR
This study evaluated citation accuracy in ChatGPT (GPT-4)-generated musculoskeletal rehabilitation texts and found that most initial references were inaccurate or fabricated. A structured post-generation verification prompt substantially improved reference accuracy, but the authors conclude that LLM-assisted writing still requires strict human-led citation verification.
Key findings
- In the initial generation, only 3/40 references (7.5%) were fully accurate, 20/40 (50%) were partially correct, and 17/40 (42.5%) were completely fabricated or had major errors.
- After ChatGPT was prompted to verify and revise its references, 31/40 (77.5%) were fully accurate; median accuracy increased from 1 (IQR 0–1) to 2 (IQR 2–2), with Wilcoxon signed-rank test W=561.0, p<0.001, r=0.60. The figure reports fabricated references declined from 42.5% to 2.5%.
- Prompt-level improvements were statistically significant: manual therapy W=32.5, p=0.017; ACL reconstruction W=45.0, p=0.003; low back pain W=48.0, p=0.019; rotator cuff repair W=34.5, p=0.008.
- Common errors included invalid DOIs, fabricated article titles, and mismatched metadata, often combining real author names with invented bibliographic details. Baseline accuracy was lowest for rotator cuff repair, where only 1/10 references (10%) was fully accurate initially.
Methods
- Structured observational analysis using four ChatGPT (GPT-4) prompts on manual therapy, ACL reconstruction, low back pain, and rotator cuff repair; each prompt requested a 300–400 word scientific paragraph with 10 references including DOIs, yielding 40 references. Prompts were administered sequentially in fixed order in separate, context-cleared sessions via the OpenAI web interface in June 2025.
- References were scored 0=fabricated or completely mismatched, 1=partially correct, 2=fully accurate. One researcher designed prompts and generated content; another independently evaluated references and DOIs, with manual verification via Crossref, PubMed, Google Scholar, and publisher websites when DOIs did not resolve.
- Descriptive statistics compared score frequencies before and after a verification prompt; Wilcoxon signed-rank tests assessed changes in accuracy, with effect sizes (r) calculated. Statistical significance was set at p<0.05, using IBM SPSS Statistics 25.
Limitations
- The paper states that the sample was limited to 40 references across four prompts, which may limit generalizability; the prompts do not cover the full scope of musculoskeletal rehabilitation.
- The paper states that reference accuracy was assessed manually and remains subject to interpretation bias despite predefined criteria, and that the verification step relied on ChatGPT’s own response to follow-up prompts rather than external validation.
- The paper states that only GPT-4 was tested, so findings may differ across other models or platforms. The text also makes evident that prompts were not randomized and that the evaluator’s blinding to condition or verification status is not reported.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
12 Valid · 0 Not valid · 10 Not checked · 79 Own results and interpretations
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“In contrast to previous anecdotal reports or general evaluations, 20,21 there remains a lack of systematic research quantifying the accuracy and correction potential of hallucinated references in musculoskeletal-focused scientific content.”
What the sources say
- Supportedcites [20] Gravel J, D’Amours-Gravel M, Osmanlliu E. Learning to fake it: limited responses and fabricated references provided by ChatGPT for medical questions. Mayo Clin Proc Digit Health. 2023;1(3):226-234. doi:10.1016/j.mcpdig.2023.05.004 source
The citation supports the sentence’s reference to previous general evaluations: the source’s Patients and Methods section describes a study measuring the validity of ChatGPT-generated references across varied medical questions. The claimed gap concerning musculoskeletal content and correction potential is the citing authors’ own framing, not a finding attributed to this source. Page numbers were not available in our copy. (LLM, conf=0.88)
- Not checked — no full text of the sourcecites [21] Orduña-Malea E, Cabezas-Clavijo Á. ChatGPT and the potential growing of ghost bibliographic references. Scientometrics. 2023;128(9): 5351-5365. doi:10.1007/s11192-023-04804-4 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“Our results align with previous research demonstrating significant challenges associated with fabricated and inaccurate references produced by ChatGPT. 22-24”
What the sources say
- Not checked — no full text of the sourcecites [22] Lechien JR, Briganti G, Vaira LA. Accuracy of ChatGPT-3.5 and -4 in providing scientific references in otolaryngology–head and neck surgery. Eur Arch Otorhinolaryngol. 2024;281(4):2159-2165. doi:10.1007/s00405-023-08441-8 source
- Supportedcites [23] Sanchez-Ramos L, Lin L, Romero R. Beware of references when using ChatGPT as a source of information to write scientific articles. Am J Obstet Gynecol. 2023;229(3):356. doi:10.1016/j.ajog.2023.04.004 source
The letter (p. 1) states that the authors found ChatGPT frequently produced erroneous and fictitious references, including incorrect authors, article titles, publication years, and identifiers. It also says such errors undermine a scientific report’s credibility, supporting the claim that they pose significant challenges. (LLM, conf=0.94)
- Not checked — no full text of the sourcecites [24] Sebo P. How accurate are the references generated by ChatGPT in internal medicine? Intern Emerg Med. 2024;19(1):247-249. doi:10.1007/s11739-023-03484-5 source
boostedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Common errors included incorrect PMIDs (93%), volumes, pages, and publication years.”
What the sources say
- Partly supportedcites [33] Bhattacharyya M, Miller VM, Bhattacharyya D, Miller LE, Miller V. High rates of fabricated and inaccurate references in ChatGPT-generated medical content. Cureus. 2023;15(5):e39238. doi:10.7759/cureus.39238 sourceExperimental — accuracy not yet measured
The Discussion (page unavailable) says that over 90% of references had an incorrect PMID, but does not give the precise 93% figure for PMIDs; its 93% figure refers to references with any inaccuracy. Materials and methods (page unavailable) lists volume, pages, and year as elements assessed, but the retrieved passages do not establish that errors in each were common. (LLM, conf=0.84)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“When references generated by LLMs are inaccurate or fabricated, the consequences extend beyond academic integrity, posing risks for patient safety, clinical outcomes, and professional training.”
What the sources say
- Partly supportedcites [5] Švab I, Klemenc-Ketiš Z, Zupanič S. New challenges in scientific publications: referencing, artificial intelligence and ChatGPT. Slov J Public Health. 2023;62(3):109-112. doi:10.2478/sjph-2023-0015 sourceExperimental — accuracy not yet measured
The AI-powered research tools section (page not specified) says ChatGPT can make factual errors, generate nonexistent references, and mislead authors and readers. It does not link those errors to patient safety, clinical outcomes, or professional training, so the claimed consequences are not supported by this paper. (LLM, conf=0.86)
- Not checked — no full text of the sourcecites [6] Jan R. Examining the reliability of ChatGPT: identifying retracted scientific literature and ensuring accurate citations and references. In: Impacts of generative ai on the future of research and education. Hershey, PA: IGI Global; 2025. source
- Not checked — no full text of the sourcecites [7] Frosolini A, Gennaro P, Cascino F, Gabriele G. In reference to “role of Chat GPT in public health”, to highlight the AI’s incorrect reference generation. Ann Biomed Eng. 2023;51(10):2120-2122. doi:10.1007/s 10439-023-03248-4 source
- Not checked — no full text of the sourcecites [8] Cohen F, Vallimont J, Gelfand AA. Caution regarding fabricated citations from artificial intelligence. Headache. 2024;64(1):133-135. doi: 10.1111/head.14649 source
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Recent investigations have increasingly drawn attention to the phenomenon of 'hallucinated references,' wherein LLMs such as ChatGPT generate citations that appear syntactically correct and contextually appropriate, yet do not correspond to any real or retrievable publication.”
What the sources say
- Partly supportedcites [9] Borji A. A categorical archive of ChatGPT failures. arXiv. 2023;2302. 03494. doi:10.48550/arXiv.2302.03494 sourceExperimental — accuracy not yet measured
Section 2.4, Factual Errors (p. 12), says ChatGPT can produce credible-seeming but inaccurate information and can fabricate facts. It does not specifically describe generated citations that look valid yet refer to no real or retrievable publication. (LLM, conf=0.81)
- Partly supportedcites [10] Bai Z, Wang P, Xiao T, et al. Hallucination of multimodal large language models: a survey. arXiv. 2024;2404.18930. doi:10.48550/arXiv.2404.18930 sourceExperimental — accuracy not yet measured
Our copy supports only the generic layer of the claim: the Introduction (p. 2) states that LLMs "exhibit an inclination to generate hallucinations, resulting in seemingly plausible yet factually spurious content," and categorises factuality hallucination as "the discrepancy between generated content and verifiable real-world facts, typically manifesting as factual inconsistency or fabrication" (p. 2). Nothing in the survey addresses the specific phenomenon the citing sentence attributes to it — LLMs such as ChatGPT producing hallucinated references/citations that look syntactically and contextually valid but correspond to no real publication. Searched for 'hallucinated citations', 'fabricated/non-existent references', 'bibliography', 'fake publication'; every matched passage on fabrication concerns cross-modal object/event hallucination in MLLMs (e.g. Hal-Eval, p. 13; NOPE, p. 11), not bibliographic records. The ChatGPT mentions in our copy (HaELM p. 12, Volcano p. 27, GAVIE p. 11) concern its use as evaluator/data organiser, not citation generation. (LLM, conf=0.66)
- Not checked — no full text of the sourcecites [11] Metze K, Morandin-Reis RC, Lorand-Metze I, Florindo JB. Bibliographic research with ChatGPT may be misleading: the problem of hallucination. J Pediatr Surg. 2024;59(1):158-159. doi:10.1016/j.jpedsurg.2023.08.018 source
- Supportedcites [12] Aljamaan F, Temsah M-H, Altamimi I, et al. Reference hallucination score for medical artificial intelligence chatbots: development and usability study. JMIR Med Inform. 2024;12(1):e54345. doi:10.2196/54345 source
The Introduction (page unavailable) says ChatGPT and other chatbots generated citations with detailed bibliographic data that seemed authentic, but many were fabricated or erroneous when checked against medical literature resources. It identifies these fictitious citations as hallucinations or fabrications. (LLM, conf=0.90)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
56 more claims · 1 with their cited sources checked
Create a free account to see all 101 claims and every source verdict.
- Probably not a substantial claim. 23
- It may state two separate claims as one. 27
Citations
Cites
Tufford (2017) · doi:10.1007/s · Cited by 2,003
Artificial hallucinations in ChatGPT: implications in scientific writing
Alkaissi, H., & McFarlane, S. I. (2023) · Cureus · doi:10.7759/cureus.35179 · Cited by 1,540
ChatGPT is fun, but not an author
Thorp, H. H. (2023) · Science · doi:10.1126/science.adg7879 · Cited by 1,437
Gao, C. A., Howard, F. M., Markov, N. S., Dyer, E. C., Ramesh, S., Luo, Y., & Pearson, A. T. (2022) · doi:10.1101/2022.12.23.521610 · Cited by 1,099
Gao, C. A., Howard, F. M., Markov, N. S., Dyer, E. C., Ramesh, S., Luo, Y., & Pearson, A. T. (2023) · npj Digital Medicine · doi:10.1038/s41746-023-00819-6 · Cited by 1,099
A Categorical Archive of ChatGPT Failures
Borji, A. (2023) · doi:10.21203/rs.3.rs-2895792/v1 · Cited by 514
A Categorical Archive of ChatGPT Failures
Borji, A. (2023) · arXiv · doi:10.48550/arxiv.2302.03494 · Cited by 514
Fabrication and errors in the bibliographic citations generated by ChatGPT
Walters, W. H., & Wilder, E. I. (2023) · Scientific Reports · doi:10.1038/s41598-023-41032-5 · Cited by 499
Athaluri, S. A., Manthena, S. V., Kesapragada, V. S. R. K. M., Yarlagadda, V., Dave, T., & Duddumpudi, R. T. S. (2023) · Cureus · doi:10.7759/cureus.37432 · Cited by 403
High Rates of Fabricated and Inaccurate References in ChatGPT-Generated Medical Content
Bhattacharyya, M., Miller, V. M., Bhattacharyya, D., & Miller, L. E. (2023) · Cureus · doi:10.7759/cureus.39238 · Cited by 294
Cited by
Pratama, R. D., Widiati, U., Basthomi, Y., Khoiri, N. E., Muniroh, S., & Omolu, F. A. (2026) · Journal of Academic Ethics · doi:10.1007/s10805-026-09754-z · Cited by 2
“Although previous literature has high - lighted the benefits of AI tools in improving the efficiency and quality of writing, as well as the potential of the HAICW strategy in strengthening students’ critical thinking skills and reducing AI-hallucinations and AI overreliance, studies linking it to aspects of ethics and…”
InfluentialKhan, H., Nafis, M. T., Sohail, S. S., Khalique, A., & Khan, R. H. (2026) · International Conference on Emerging Computational Intelligence (ICECI) · doi:10.1109/iceci69159.2026.11519401 · Cited by 1
Rathomi, A., Marjuni, & Afifudin (2026) · SIWAYANG JOURNAL PUBLIKASI ILMIAH BIDANG PARIWISATA KEBUDAYAAN DAN ANTROPOLOGI · doi:10.47353/siwayang.v5i1.5233
Морозова, С. А. (2026) · Bibliotekovedenie [Russian Journal of Library Science] · doi:10.25281/0869-608x-2026-75-4-355-370
Metadata
- Authors
- Safran, E., & Çalı, A.
- Affiliations
- Bezmiâlem Vakıf Üniversitesi (TR)
- Istanbul Topkapi University (TR)
- Venue
- Anatolian Current Medical Journal
- Publisher
- Anatolian Current Medical Journal
- Published
- 2025-09-15
- Volume
- 7
- Issue
- 5
- Pages
- 695-702
- ISSN
- 2718-0115 (ISSN-L)
- OpenAlex
- W4414428938
- Cited by
- 4
- References
- 32
- Citations by year
- 2026: 4
- FWCI
- 0.97
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by-nc-nd
- Locations
- Anatolian Current Medical Journal (publishedVersion, cc-by-nc-nd)
- DergiPark (Istanbul University) (submittedVersion)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Social Sciences › Library and Information Sciences › Academic Writing and Publishing
- Social Sciences › Education › Academic integrity and plagiarism
- Keywords
- citation hallucination, ChatGPT, large language models, citation accuracy, musculoskeletal rehabilitation, manual therapy, anterior cruciate ligament reconstruction, low back pain, rotator cuff repair, research ethics
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 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
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: Fabricated or accurate? Ethical concerns and citation hallucination in aI-generated scientific writing on musculoskeletal topics. https://app.cochat.ai/papers/10.38053/acmj.1746227
- The paper
- Safran, E., & Çalı, A. (2025). Fabricated or accurate? Ethical concerns and citation hallucination in aI-generated scientific writing on musculoskeletal topics. Anatolian Current Medical Journal. https://doi.org/10.38053/acmj.1746227
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.