{"title":"对ChatGPT可读性的评估产生了关于旅行健康风险的回应。","authors":"Zeynep Demirtaş Özyılmaz","doi":"10.5603/imh.104400","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Impressive advances in artificial intelligence have given travellers a new source of health information. AI-powered tools, such as ChatGPT, allow users to obtain health information in a fast and accessible way. The aim of this study was to assess the readability of ChatGPT responses to questions about health risks when travelling.</p><p><strong>Material and methods: </strong>Ten Questions about the health risks when travelling in the \"Questions and Answers\" section of the World Health Organisation's website have been asked 10 times to ChatGPT. A total of 100 answers was obtained and analyzed for readability.</p><p><strong>Results: </strong>The mean ± SD of Flesch Reading Ease was 35.82 ± 6.46, Flesch-Kincaid grade level was 13.25 ± 1.45, Simple Measure of Gobbledygook was 12.34 ± 1.29, Gunning Fog Index was 13.77 ± 1.34, Coleman-Liau Index was 14.52 ± 1.09, Automated Readability Index was 14.93 ± 1.81.</p><p><strong>Conclusions: </strong>The readability of the answers produced by ChatGPT was 'difficult' and a college level education is required to understand the text. Lack of understanding of information can reduce the likelihood of travellers making good health decisions. To improve the understandability of ChatGPT responses, it may be useful to generate responses at a significantly lower reading level.</p>","PeriodicalId":45964,"journal":{"name":"International Maritime Health","volume":" ","pages":""},"PeriodicalIF":1.0000,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Evaluating of the readability of ChatGPT generated responses on travel health risks.\",\"authors\":\"Zeynep Demirtaş Özyılmaz\",\"doi\":\"10.5603/imh.104400\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Background: </strong>Impressive advances in artificial intelligence have given travellers a new source of health information. AI-powered tools, such as ChatGPT, allow users to obtain health information in a fast and accessible way. The aim of this study was to assess the readability of ChatGPT responses to questions about health risks when travelling.</p><p><strong>Material and methods: </strong>Ten Questions about the health risks when travelling in the \\\"Questions and Answers\\\" section of the World Health Organisation's website have been asked 10 times to ChatGPT. A total of 100 answers was obtained and analyzed for readability.</p><p><strong>Results: </strong>The mean ± SD of Flesch Reading Ease was 35.82 ± 6.46, Flesch-Kincaid grade level was 13.25 ± 1.45, Simple Measure of Gobbledygook was 12.34 ± 1.29, Gunning Fog Index was 13.77 ± 1.34, Coleman-Liau Index was 14.52 ± 1.09, Automated Readability Index was 14.93 ± 1.81.</p><p><strong>Conclusions: </strong>The readability of the answers produced by ChatGPT was 'difficult' and a college level education is required to understand the text. Lack of understanding of information can reduce the likelihood of travellers making good health decisions. To improve the understandability of ChatGPT responses, it may be useful to generate responses at a significantly lower reading level.</p>\",\"PeriodicalId\":45964,\"journal\":{\"name\":\"International Maritime Health\",\"volume\":\" \",\"pages\":\"\"},\"PeriodicalIF\":1.0000,\"publicationDate\":\"2025-09-16\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Maritime Health\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.5603/imh.104400\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Maritime Health","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.5603/imh.104400","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH","Score":null,"Total":0}
引用次数: 0
摘要
背景：人工智能令人印象深刻的进步为旅行者提供了新的健康信息来源。ChatGPT等人工智能工具允许用户以快速和可访问的方式获取健康信息。本研究的目的是评估ChatGPT对旅行时健康风险问题的回答的可读性。资料和方法：在世界卫生组织网站的“问答”部分，关于旅行时的健康风险的十个问题已经被问了十次。共获得100个答案，并对其可读性进行分析。结果：Flesch Reading Ease的平均值±SD为35.82±6.46，Flesch- kincaid分级水平为13.25±1.45，Simple Measure of Gobbledygook为12.34±1.29，Gunning Fog指数为13.77±1.34，Coleman-Liau指数为14.52±1.09，Automated Readability指数为14.93±1.81。结论：ChatGPT产生的答案的可读性是“困难的”，并且需要大学水平的教育才能理解文本。缺乏对信息的了解会降低旅行者作出良好健康决定的可能性。为了提高ChatGPT响应的可理解性，在较低的阅读水平下生成响应可能是有用的。
Evaluating of the readability of ChatGPT generated responses on travel health risks.
Background: Impressive advances in artificial intelligence have given travellers a new source of health information. AI-powered tools, such as ChatGPT, allow users to obtain health information in a fast and accessible way. The aim of this study was to assess the readability of ChatGPT responses to questions about health risks when travelling.
Material and methods: Ten Questions about the health risks when travelling in the "Questions and Answers" section of the World Health Organisation's website have been asked 10 times to ChatGPT. A total of 100 answers was obtained and analyzed for readability.
Results: The mean ± SD of Flesch Reading Ease was 35.82 ± 6.46, Flesch-Kincaid grade level was 13.25 ± 1.45, Simple Measure of Gobbledygook was 12.34 ± 1.29, Gunning Fog Index was 13.77 ± 1.34, Coleman-Liau Index was 14.52 ± 1.09, Automated Readability Index was 14.93 ± 1.81.
Conclusions: The readability of the answers produced by ChatGPT was 'difficult' and a college level education is required to understand the text. Lack of understanding of information can reduce the likelihood of travellers making good health decisions. To improve the understandability of ChatGPT responses, it may be useful to generate responses at a significantly lower reading level.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
