ChatGPT- versus human-generated answers to frequently asked questions about diabetes: A Turing test-inspired survey among employees of a Danish diabetes center
Open accessThe DOI matches its Crossref record; 24 of the 24 references it cites were found in a registry or an index.
Abstract
Large language models have received enormous attention recently with some studies demonstrating their potential clinical value, despite not being trained specifically for this domain. We aimed to investigate whether ChatGPT, a language model optimized for dialogue, can answer frequently asked questions about diabetes. We conducted a closed e-survey among employees of a large Danish diabetes center. The study design was inspired by the Turing test and non-inferiority trials. Our survey included ten questions with two answers each. One of these was written by a human expert, while the other was generated by ChatGPT. Participants had the task to identify the ChatGPT-generated answer. Data was analyzed at the question-level using logistic regression with robust variance estimation with clustering at participant level. In secondary analyses, we investigated the effect of participant characteristics on the outcome. A 55% non-inferiority margin was pre-defined based on precision simulations and had been published as part of the study protocol before data collection began. Among 311 invited individuals, 183 participated in the survey (59% response rate). 64% had heard of ChatGPT before, and 19% had tried it. Overall, participants could identify ChatGPT-generated answers 59.5% (95% CI: 57.0, 62.0) of the time, which was outside of the non-inferiority zone. Among participant characteristics, previous ChatGPT use had the strongest association with the outcome (odds ratio: 1.52 (1.16, 2.00), p = 0.003). Previous users answered 67.4% (61.7, 72.7) of the questions correctly, versus non-users’ 57.6% (54.9, 60.3). Participants could distinguish between ChatGPT-generated and human-written answers somewhat better than flipping a fair coin, which was against our initial hypothesis. Rigorously planned studies are needed to elucidate the risks and benefits of integrating such technologies in routine clinical practice.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
Eysenbach, G. (2004) · Journal of Medical Internet Research · doi:10.2196/jmir.6.3.e34 · Cited by 6,480
Kung, T. H., Cheatham, M., Medenilla, A., Sillos, C., De Leon, L., Elepaño, C., Madriaga, M., Aggabao, R., Diaz-Candido, G., Maningo, J., et al. (2023) · PLOS Digital Health · doi:10.1371/journal.pdig.0000198 · Cited by 4,115 · cited 2×
“More studies and the scientific discourse have mostly focused on ChatGPT’s role in medical research (e.g. scientific writing 6,19,20 ), formal testing of large language models’ medical knowledge e.g. by taking medical licensing exams, 7-9 or answering questions in highly specialized topics.”
Kung, T. H., Cheatham, M., ChatGPT, Medenilla, A., Sillos, C., De Leon, L., Elepaño, C., Madriaga, M., Aggabao, R., Diaz-Candido, G., et al. (2022) · doi:10.1101/2022.12.19.22283643 · Cited by 4,115
Large Language Models Encode Clinical Knowledge
Singhal, K., Azizi, S., Tu, T., Mahdavi, S. S., Wei, J., Chung, H. W., Scales, N., Tanwani, A., Cole-Lewis, H., Pfohl, S., et al. (2022) · arXiv · doi:10.48550/arxiv.2212.13138 · Cited by 4,083
Large language models encode clinical knowledge
Singhal, K., Azizi, S., Tu, T., Mahdavi, S. S., Wei, J., Chung, H. W., Scales, N., Tanwani, A., Cole-Lewis, H., Pfohl, S., et al. (2023) · Nature · doi:10.1038/s41586-023-06291-2 · Cited by 4,083 · cited 1×
“More studies and the scientific discourse have mostly focused on ChatGPT’s role in medical research (e.g. scientific writing 6,19,20 ), formal testing of large language models’ medical knowledge e.g. by taking medical licensing exams, 7-9 or answering questions in highly specialized topics.”
Principles alone cannot guarantee ethical AI
Mittelstadt, B. (2019) · Nature Machine Intelligence · doi:10.1038/s42256-019-0114-4 · Cited by 1,886 · cited 1×
“28,29 One feature that could strengthen trust would be if large language models provided trustworthy references supporting their answers.”
ChatGPT: five priorities for research
van Dis, E. A. M., Bollen, J., Zuidema, W., van Rooij, R., & Bockting, C. L. (2023) · Nature · doi:10.1038/d41586-023-00288-7 · Cited by 1,842
ChatGPT is fun, but not an author
Thorp, H. H. (2023) · Science · doi:10.1126/science.adg7879 · Cited by 1,437
Exercise management in type 1 diabetes: a consensus statement
Riddell, M. C., Gallen, I. W., Smart, C. E., Taplin, C. E., Adolfsson, P., Lumb, A. N., Kowalski, A., Rabasa-Lhoret, R., McCrimmon, R. J., Hume, C., et al. (2017) · The Lancet Diabetes & Endocrinology · doi:10.1016/s2213-8587(17)30014-1 · Cited by 903
Morley, J., Floridi, L., Kinsey, L., & Elhalal, A. (2020) · Science and Engineering Ethics · doi:10.1007/s11948-019-00165-5 · Cited by 882
ChatGPT listed as author on research papers: many scientists disapprove
Stokel-Walker, C. (2023) · Nature · doi:10.1038/d41586-023-00107-z · Cited by 868
Generating scholarly content with ChatGPT: Ethical challenges for medical publishing
Liebrenz, M., Schleifer, R., Buadze, A., Bhugra, D., & Smith, A. (2023) · The Lancet Digital Health · doi:10.1016/s2589-7500(23)00019-5 · Cited by 594 · cited 1×
“More studies and the scientific discourse have mostly focused on ChatGPT’s role in medical research (e.g. scientific writing 6,19,20 ), formal testing of large language models’ medical knowledge e.g. by taking medical licensing exams, 7-9 or answering questions in highly specialized topics.”
Evaluating the Performance of ChatGPT in Ophthalmology
Antaki, F., Touma, S., Milad, D., El-Khoury, J., & Duval, R. (2023) · Ophthalmology Science · doi:10.1016/j.xops.2023.100324 · Cited by 563 · cited 1×
“10,11 A potential role of ChatGPT in chronic disease management, like diabetes, is yet to be debated.”
Laux, J., Wachter, S., & Mittelstadt, B. (2024) · Regulation & Governance · doi:10.1111/rego.12512 · Cited by 365
Can large language models reason about medical questions?
Liévin, V., Hother, C. E., Motzfeldt, A. G., & Winther, O. (2024) · Patterns · doi:10.1016/j.patter.2024.100943 · Cited by 343
Can large language models reason about medical questions?
Liévin, V., Hother, C. E., Motzfeldt, A. G., & Winther, O. (2022) · arXiv · doi:10.48550/arxiv.2207.08143 · Cited by 343
Health, T. L. D. (2023) · The Lancet Digital Health · doi:10.1016/s2589-7500(23)00023-7 · Cited by 257
Atlas: Few-shot Learning with Retrieval Augmented Language Models
Izacard, G., Lewis, P., Lomeli, M., Hosseini, L., Petroni, F., Schick, T., Dwivedi-Yu, J., Joulin, A., Riedel, S., & Grave, E. (2022) · arXiv · doi:10.48550/arxiv.2208.03299 · Cited by 213
Putting ChatGPT’s Medical Advice to the (Turing) Test: Survey Study
Nov, O., Singh, N., & Mann, D. (2023) · JMIR Medical Education · doi:10.2196/46939 · Cited by 198
Analysis of large-language model versus human performance for genetics questions
Duong, D., & Solomon, B. D. (2023) · doi:10.1101/2023.01.27.23285115 · Cited by 159
Analysis of large-language model versus human performance for genetics questions
Duong, D., & Solomon, B. D. (2024) · European Journal of Human Genetics · doi:10.1038/s41431-023-01396-8 · Cited by 159
Smuha, N. A., Rengers, E., Harkens, A., Li, W., MacLaren, J., Piselli, R., & Yeung, K. (2021) · doi:10.2139/ssrn.3899991 · Cited by 153 · cited 1×
“…when it comes to AI” 24 , has been touted by EU officials as addressing concerns around large language model-generated misinformation, but has been criticized for putting the onus on AI developers to undertake risk assessment, 25 and for conflating risk acceptability with trustworthiness.”
Caution with AI-generated content in biomedicine
Zhavoronkov, A. (2023) · Nature Medicine · doi:10.1038/d41591-023-00014-w · Cited by 39
Cited by
Nikolic, S., Daniel, S., Haque, R., Belkina, M., Hassan, G. M., Grundy, S., Lyden, S., Neal, P., & Sandison, C. (2023) · European Journal of Engineering Education · doi:10.1080/03043797.2023.2213169 · Cited by 336
“In two recent studies in a medical context, ChatGPT’s responses were only weakly distinguishable from those generated by a real person (Hulman et al. 2023; Nov, Singh, and Mann 2023).”
BackgroundA Survey of GPT-3 Family Large Language Models Including ChatGPT and GPT-4
Kalyan, K. S. (2024) · Natural Language Processing Journal · doi:10.1016/j.nlp.2023.100048 · Cited by 335
Moulaei, K., Yadegari, A., Baharestani, M., Farzanbakhsh, S., Sabet, B., & Reza Afrash, M. (2024) · International Journal of Medical Informatics · doi:10.1016/j.ijmedinf.2024.105474 · Cited by 244
Let's have a chat! A Conversation with ChatGPT: Technology, Applications, and Limitations
Shahriar, S., & Hayawi, K. (2023) · arXiv · doi:10.48550/arxiv.2302.13817 · Cited by 167
Let’s Have a Chat! A Conversation with ChatGPT: Technology, Applications, and Limitations
Shahriar, S., & Hayawi, K. (2023) · Artificial Intelligence and Applications · doi:10.47852/bonviewaia3202939 · Cited by 167
“Improving efficiency in Urology (Gabrielson et al., 2023) Discussed the role of generative AI in helping urologists N/A Answer frequently asked medical questions (Hulman et al., 2023) Asked human evaluators to distinguish between ChatGPT and human answers to diabetes-related questions The evaluators could only identif…”
BackgroundInfluential
Metadata
- Authors
- Hulman, A., Dollerup, O. L., Mortensen, J. F., Fenech, M. E., Norman, K., Støvring, H., & Hansen, T. K.
- Venue
- PLOS ONE
- Publisher
- Public Library of Science (PLoS)
- Published
- 2023-08-31
- Volume
- 18
- Issue
- 8
- Pages
- e0290773
- ISSN
- 1932-6203 (ISSN-L)
- OpenAlex
- W4386307074
- Cited by
- 70
- References
- 24
- FWCI
- 2.42
- Type
- journal article
- Language
- en
- Open access
- gold
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
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.