A comparative evaluatıon of AI-based ChatGPT and physicians in psychiatric case assessment

人工智能聊天机器人与医生在精神科病例评估中的比较评价

阅读:1

Abstract

According to the results of our study, ChatGPT shows a variable performance in different aspects in medical fields. It was observed that it did not reach the level of physicians in communicating with the patient and conducting in-depth questioning. However, in most cases, the individuals' past psychiatric illnesses or medical histories are questioned. However, especially in cases where risk assessment is required, this process does not achieve the expected level of success and lacks the depth required for an effective assessment. While it functions effectively to inform differential diagnoses, its ability to make complex clinical decisions and make definitive diagnoses is limited. While it generally performed adequately in determining which medications should be prescribed and appropriate dosages, its capacity to warn about side effects of medications was found to be relatively weak. ChatGPT, which performs strongly in providing supportive treatment recommendations, shows a marked inability in more complex clinical processes such as making patient hospitalization decisions. These results suggest that AI-based systems can be useful as assistive tools in medical practice, but should be used with an awareness of their limitations.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。