ArticlePloS one2024
Assessing the accuracy and consistency of answers by ChatGPT to questions regarding carbon monoxide poisoning.
Article in PloS one, 2024. The graph could read no effect estimate from its abstract, so it casts no vote on the map. Cited by 1 paper.
What it found
Each row is one number read from the abstract, on the scale the paper reported it, with its interval. Left of the dashed line favours the treatment, right favours the comparator. Under each row is the sentence it came from. New to these charts? A ten-minute tutorial.
The abstract states no effect estimate the extractor could read, or names no intervention and outcome on the map, so this paper lights no cell and moves no belief. It is still indexed, cited and linked below.
The trial behind it
Trials whose registry record cites this paper, or whose number appears in the abstract. A trial that started after this paper was published is citing it as background, not reporting it.
Neither the registry nor the abstract names a trial number. If this is a trial report, that itself is worth knowing.
Who cites it
1 citing paper in PubMed.
Corrections and comments
PubMed lists nothing against this paper. Absence here is not a guarantee, only a check that was made.
Authors and funding
2 authors.
Funding
No grant is acknowledged in the PubMed record.
Abstract
backgroundChatGPT, developed by OpenAI, is an artificial intelligence software designed to generate text-based responses. The objective of this study is to evaluate the accuracy and consistency of ChatGPT's responses to single-choice questions pertaining to carbon monoxide poisoning. This evaluation will contribute to our understanding of the reliability of ChatGPT-generated information in the medical field.
methodsThe questions utilized in this study were selected from the "Medical Exam Assistant (Yi Kao Bang)" application and encompassed a range of topics related to carbon monoxide poisoning. A total of 44 single-choice questions were included in the study following a screening process. Each question was entered into ChatGPT ten times in Chinese, followed by a translation into English, where it was also entered ten times. The responses generated by ChatGPT were subjected to statistical analysis with the objective of assessing their accuracy and consistency in both languages. In this assessment process, the "Medical Exam Assistant (Yi Kao Bang)" reference responses were employed as benchmarks. The data analysis was conducted using the Python.
resultsIn approximately 50% of the cases, the responses generated by ChatGPT exhibited a high degree of consistency, whereas in approximately one-third of the cases, the responses exhibited unacceptable blurring of the answers. Meanwhile, the accuracy of these responses was less favorable, with an accuracy rate of 61.1% in Chinese and 57% in English. This indicates that ChatGPT could be enhanced with respect to both consistency and accuracy in responding to queries pertaining to carbon monoxide poisoning.
conclusionsIt is currently evident that the consistency and accuracy of responses generated by ChatGPT regarding carbon monoxide poisoning is inadequate. Although it offers significant insights, it should not supersede the role of healthcare professionals in making clinical decisions.
Indexed as
Identifiers
What OpenQuestion holds
Registered trials
Read under generation 80e0d062 · epoch 390. Bibliography from PubMed, PubMed Central and OpenAlex; grants from NIH RePORTER; trial links from ClinicalTrials.gov; estimates, votes and beliefs from the OpenQuestion graph.