The DOI matches its Crossref record; 39 of the 39 references it cites were found in a registry or an index; 27 of the 33 sources its claims cite were checked for support.
Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study
Open accessPublished version of a preprint
Abstract
Background: Artificial intelligence (AI) chatbots have recently gained use in medical practice by health care practitioners. Interestingly, the output of these AI chatbots was found to have varying degrees of hallucination in content and references. Such hallucinations generate doubts about their output and their implementation. Objective: The aim of our study was to propose a reference hallucination score (RHS) to evaluate the authenticity of AI chatbots’ citations. Methods: Six AI chatbots were challenged with the same 10 medical prompts, requesting 10 references per prompt. The RHS is composed of 6 bibliographic items and the reference’s relevance to prompts’ keywords. RHS was calculated for each reference, prompt, and type of prompt (basic vs complex). The average RHS was calculated for each AI chatbot and compared across the different types of prompts and AI chatbots. Results: Bard failed to generate any references. ChatGPT 3.5 and Bing generated the highest RHS (score=11), while Elicit and SciSpace generated the lowest RHS (score=1), and Perplexity generated a middle RHS (score=7). The highest degree of hallucination was observed for reference relevancy to the prompt keywords (308/500, 61.6%), while the lowest was for reference titles (169/500, 33.8%). ChatGPT and Bing had comparable RHS (β coefficient=–0.069; P=.32), while Perplexity had significantly lower RHS than ChatGPT (β coefficient=–0.345; P<.001). AI chatbots generally had significantly higher RHS when prompted with scenarios or complex format prompts (β coefficient=0.486; P<.001). Conclusions: The variation in RHS underscores the necessity for a robust reference evaluation tool to improve the authenticity of AI chatbots. Further, the variations highlight the importance of verifying their output and citations. Elicit and SciSpace had negligible hallucination, while ChatGPT and Bing had critical hallucination levels. The proposed AI chatbots’ RHS could contribute to ongoing efforts to enhance AI’s general reliability in medical research.
Overview
CoChat Summary
TL;DR
The study proposes a Reference Hallucination Score (RHS) to evaluate the authenticity of AI chatbot citations, scoring 7 bibliographic/relevance items with a maximum score of 11. Across 6 chatbots and 500 requested references, Elicit and SciSpace had negligible hallucination, while ChatGPT 3.5 and Bing had critical hallucination and Bard generated no references.
Key findings
- Bard failed to generate any references for all 10 prompts; overall, 35/500 (7%) references were not generated. The highest hallucination was reference relevance to prompt keywords (308/500, 61.6%), while the lowest was reference title (169/500, 33.8%); publication date was 237/500 (47.4%), authors’ names 228/500 (45.6%), DOI 227/500 (45.4%), web link 187/500 (37.4%), and journal name 188/500 (37.6%).
- Median total RHS was highest for ChatGPT 3.5 and Bing (score=11) and lowest for Elicit and SciSpace (score=1), with Perplexity in the middle (score=7).
- In multivariable regression, ChatGPT and Bing had comparable RHS (β coefficient=–0.069; P=.32), while Perplexity had significantly lower RHS than ChatGPT (β coefficient=–0.345; P<.001); SciSpace (β coefficient=–1.748; P<.001) and Elicit (β coefficient=–1.630; P<.001) had the lowest scores. Complex/scenario prompts produced significantly higher RHS than basic prompts (β coefficient=0.486; P<.001).
- SciSpace and Elicit had the highest correct reference identifiers (629 and 597, respectively); ChatGPT had the highest hallucination results (592), and Bing had the highest rate of failure to generate results (210).
Methods
- RHS was developed using a literature review and Delphi technique: 2 senior librarians, 12 senior physicians, and 11 junior physicians were consulted, followed by author consensus. Seven items were scored: erroneous title, journal name, authors’ names, or DOI scored 2; erroneous publication date, web link, or relevance to prompt keywords scored 1; maximum RHS=11, minimum=0.
- Six AI chatbots (ChatGPT 3.5, Bard, Perplexity, Bing, Elicit, SciSpace) received the same 10 medical prompts (5 topics, each in basic and complex/scenario format), requesting 10 references per prompt with 7 bibliographic items. References were verified against PubMed first, then Google Scholar.
- Analysis used mean/SD or median/IQR, Kruskal-Wallis and Mann-Whitney tests, and multivariable generalized linear gamma regression for mean RHS by chatbot, prompt complexity, and prompt iteration; SPSS was used with α=.05. No institutional review board approval was required.
Limitations
- The authors state the RHS is novel and liable to future improvement; it includes a limited number of bibliographic parameters, and other researchers might consider additional identifiers crucial.
- The authors state prompt structure is a limitation because LLM output depends heavily on prompt design, possibly explaining Bard’s failure, and the medical prompts may have limited scientific scope; verification via PubMed and Google Scholar may be suboptimal, only 6 chatbots were assessed, and selection biases may affect generalizability.
- Evident from the text: the paper does not report inter-rater reliability or duplicate independent verification for the reference verification/RHS scoring process.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
15 Valid · 0 Not valid · 1 Not checked · 97 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“a minority of AI chatbot users realized early that these LLMs do not voluntarily cite credible references to authenticate their information [10-12].”
§ Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study > Introduction
What the sources say
- Partly supportedcites [10] sourceExperimental — accuracy not yet measured
The Editorial’s “Important Considerations” section warns that publicly available chatbots can fabricate references and says they should clearly cite reliable sources. It does not report that a minority of users recognized this problem early, or establish how often LLMs cite credible references without prompting. (LLM, conf=0.87)
- Not checked — no full text of the sourcecites [11] source
- Partly supportedcites [12] sourceExperimental — accuracy not yet measured
The Results report that 69 of 178 ChatGPT-generated references had invalid DOIs and 28 could not be found, supporting concern about the credibility of its references. The Materials and methods describe requesting references from ChatGPT, but the paper does not establish whether LLMs cite them voluntarily or whether only a minority of users recognized the problem early. (LLM, conf=0.84)
plain- 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.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Some investigators assessed the AI-generated medical information, including the ability to verify references and its publication date, but did not pay attention specifically to certain identifiers of references that help to verify their authenticity [18].”
§ Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study > Introduction
What the sources say
- Partly supportedcites [18] sourceExperimental — accuracy not yet measured
The Materials and methods section says the study assessed AI-generated appendicitis information and whether its sources could be verified. Our copy does not show that it assessed publication dates. The Discussion explicitly mentions fabricated Digital Object Identifier (DOI) numbers, so the claim that the investigators overlooked reference identifiers is not established by the cited paper. (LLM, conf=0.83)
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
“Additionally, AI chatbots encounter a hallucination/fabrication phenomenon recognized early in their use in health care.”
§ Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study > Discussion > Principal Findings
What the sources say
- Partly supportedcites [12] sourceExperimental — accuracy not yet measured
The Introduction (page not provided) defines AI hallucination as a convincing but fabricated answer, and the Conclusions identify it as a concern for ChatGPT in medical and scientific writing. The paper does not establish that the phenomenon was recognized early in chatbots’ use in health care; its reference to “early recognition” concerns model deterioration, not the historical recognition of hallucination. (LLM, conf=0.83)
- Supportedcites [27] source
The Editorial describes Bard fabricating an answer and citation in response to a medical research question, and says the same phenomenon had previously been seen with ChatGPT and labelled “artificial hallucinations.” Its account of the chatbots’ recent introduction supports the claim that this was recognized early in their health-care-related use. (LLM, conf=0.88)
- Not checked — no full text of the sourcecites [28] source
- Supportedcites [29] source
The Abstract describes ChatGPT as still in its infancy and examines its responses to medical topics. The Additional information section says its output mixes true and fabricated material, supporting the claim that hallucination or fabrication was recognized early in this healthcare-related use. (LLM, conf=0.86)
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
“Our findings showed variations in the RHS across different AI chatbots, ranging from almost null for SciSpace and Elicit to a critically high degree of hallucination for ChatGPT [25].”
§ Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study > Discussion > Principal Findings
What the sources say
- Partly supportedcites [25] sourceExperimental — accuracy not yet measured
The Discussion (page unavailable) reports that 29%–33% of ChatGPT-generated references did not exist, supporting the claim that ChatGPT produced substantial reference hallucinations. The Methods and Results (pages unavailable) compare only earlier and updated versions of that chatbot; they do not report an RHS or findings for SciSpace and Elicit. (LLM, conf=0.87)
plainstates an association- 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
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Bing and Perplexity had no significant difference in their hallucination results (P>.99)”
§ Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study > Results
plain- 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.
50 more claims
Create a free account to see all 113 claims and every source verdict.
- It may state two separate claims as one. 32
- Probably not a substantial claim. 37
- The quote only partly states this claim. 3
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 source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“These fabrications exploit two reviewer limitations: the difficulty of verifying every citation under time pressure, and the tendency to assume that specific, technical references must be legitimate.20”
From When a Human‐in‐the‐Loop Amplifies the Risk of Misalignment: Considerations for Research Ethics Oversight (2026) · cites this paper as [20]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“including fabricated or inaccurate references in systematic reviews [7,9]”
From Retrieve-Then-Verify for Evaluating Evidence Support and Hallucination in Large Language Model–Generated Medical Information: Empirical Study (2026) · cites this paper as [7]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“This inconsistency reflects a known and well-documented limitation of LLMs: their tendency to generate fluent, plausible, and contextually appropriate outputs that nonetheless do not correspond to information explicitly present in source materials.”
From Retrieve-Then-Verify for Evaluating Evidence Support and Hallucination in Large Language Model–Generated Medical Information: Empirical Study (2026) · cites this paper as [7]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Prior work has found hallucinations in several relevant domains, including fabricated or inaccurate references in systematic reviews [7,9], unsupported claims in clinical text summarization [8], and adversarial exploitation of false clinical details in decision support systems [11].”
From Retrieve-Then-Verify for Evaluating Evidence Support and Hallucination in Large Language Model–Generated Medical Information: Empirical Study (2026) · cites this paper as [7]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing 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.”
From Fabricated or accurate? Ethical concerns and citation hallucination in aI-generated scientific writing on musculoskeletal topics (2025) · cites this paper as [12]
3 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study > Methods > Statistical Analysis
“The association between the mean RHS and tested predictor independent variables was expressed as an exponentiated β coefficient (risk rate) with its associated 95% CI.”
§ Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study > Results
“| Chatbot vs SciSpace | –1.748 (–1.880 to –1.617) | <.001 |”
Citations
Cites
I.—COMPUTING MACHINERY AND INTELLIGENCE
TURING, A. M. (1950) · Mind · doi:10.1093/mind/lix.236.433 · Cited by 10,290
Survey of Hallucination in Natural Language Generation
Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P. (2023) · ACM Computing Surveys · doi:10.1145/3571730 · Cited by 5,025 · cited 1×
“Another possible mechanism related to AI chatbots’ natural language processing methodology involves encoding and decoding defects during prompt processing, generating errors, and fabricating results [34].”
BackgroundArtificial hallucinations in ChatGPT: implications in scientific writing
Alkaissi, H., & McFarlane, S. I. (2023) · Cureus · doi:10.7759/cureus.35179 · Cited by 1,540 · cited 2×
“LLMs may deal with citations and their bibliographic identifiers as text, making them vulnerable to paraphrasing and other linguistic manipulations and perhaps as a watermark so that the output that is generated by AI could be identified versus that produced by humans to reduce AI misuse [29,31,32].”
BackgroundPreiksaitis, C., & Rose, C. (2023) · JMIR Medical Education · doi:10.2196/48785 · Cited by 552 · cited 1×
“Additional factors might include insufficient training data for the LLM [17,35,36], context misinterpretation, lack of external validation, and overreliance on pattern recognition, all of which could contribute to variable degrees of hallucination, depending on the specific LLM and prompt structure.”
BackgroundFabrication 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 · cited 3×
“Walters and Wilder [16] described many hallucinations in ChatGPT’s bibliographic citations.”
BackgroundRao, A., Pang, M., Kim, J., Kamineni, M., Lie, W., Prasad, A. K., Landman, A., Dreyer, K., & Succi, M. D. (2023) · Journal of Medical Internet Research · doi:10.2196/48659 · Cited by 412 · cited 1×
“As such, hallucinations put the whole AI chatbot data output under significant question, especially in implementing and applying AI aid into clinical practice [44,45].”
MethodAthaluri, 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 · cited 3×
“However, a critical analysis of the literature reveals that a minority of AI chatbot users realized early that these LLMs do not voluntarily cite credible references to authenticate their information [10-12].”
BackgroundTrust and medical AI: the challenges we face and the expertise needed to overcome them
Quinn, T. P., Senadeera, M., Jacobs, S., Coghlan, S., & Le, V. (2021) · Journal of the American Medical Informatics Association · doi:10.1093/jamia/ocaa268 · Cited by 287
Academic integrity and artificial intelligence: is ChatGPT hype, hero or heresy?
Currie, G. M. (2023) · Seminars in Nuclear Medicine · doi:10.1053/j.semnuclmed.2023.04.008 · Cited by 283
Gravel, J., D’Amours-Gravel, M., & Osmanlliu, E. (2023) · Mayo Clinic Proceedings: Digital Health · doi:10.1016/j.mcpdig.2023.05.004 · Cited by 222 · cited 2×
“Even when ChatGPT and other AI chatbots were challenged to authenticate their outputs with references [13,14], they generated multiple citations with detailed bibliographic data that seemed perfectly authentic, but most were actually fabricated, contained at least one falsified citation data, or…”
Background
Cited by
Temsah, A., Alhasan, K., Altamimi, I., Jamal, A., Al-Eyadhy, A., Malki, K. H., & Temsah, M.-H. (2025) · Cureus · doi:10.7759/cureus.79221 · Cited by 112
“Studies have highlighted issues such as inherent biases and the phenomenon of "hallucinations," where models generate plausible but incorrect or nonsensical information [6,7].”
Self-citationTriola, M. M., & Rodman, A. (2025) · Academic Medicine · doi:10.1097/acm.0000000000005963 · Cited by 92
Chow, J. C. L., & Li, K. (2025) · Information · doi:10.3390/info16070549 · Cited by 84
Temsah, M.-H., Alhuzaimi, A. N., Almansour, M., Aljamaan, F., Alhasan, K., Batarfi, M. A., Altamimi, I., Alharbi, A., Alsuhaibani, A. A., Alwakeel, L., et al. (2024) · Journal of Medical Systems · doi:10.1007/s10916-024-02072-0 · Cited by 66
Self-citationYang, H., Li, M., Zhou, H., Xiao, Y., Fang, Q., Zhou, S., & Zhang, R. (2025) · Journal of Medical Internet Research · doi:10.2196/70080 · Cited by 83
“Besides, though many LLMs are available, determining an optimal LLM with superior performance in arbitrary medical question types remains nontrivial, as the performance may vary significantly with different network architectures, training approaches, training corpora, and test question types [33].”
Metadata
- Authors
- Aljamaan, F., Temsah, M.-H., Altamimi, I., Al-Eyadhy, A., Jamal, A., Alhasan, K., Mesallam, T. A., Farahat, M., & Malki, K. H.
- Affiliations
- King Saud University (SA)
- Corresponding author
- Malki, Khalid H
- ORCID
- Venue
- JMIR Medical Informatics
- Publisher
- JMIR Publications Inc.
- Published
- 2024-07-31
- Volume
- 12
- Pages
- e54345
- ISSN
- 2291-9694 (ISSN-L)
- PMC
- PMC11325115
- OpenAlex
- W4400270729
- Cited by
- 211
- References
- 37
- Citations by year
- 2026: 130 · 2025: 70 · 2024: 11
- FWCI
- 17.16
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- Locations
- JMIR Medical Informatics (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
- Keywords
- AI chatbots, hallucinations, citation evaluation, bibliographic references, ChatGPT, Bing Chat, perplexity, Elicit, SciSpace, prompt complexity
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 2026-10-30, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-30 (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: Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study. https://app.cochat.ai/papers/10.2196/54345
- The paper
- Aljamaan, F., Temsah, M.-H., Altamimi, I., Al-Eyadhy, A., Jamal, A., Alhasan, K., Mesallam, T. A., Farahat, M., & Malki, K. H. (2024). Reference Hallucination Score for Medical Artificial Intelligence Chatbots: Development and Usability Study. JMIR Medical Informatics. https://doi.org/10.2196/54345
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.