Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
Prompt Design Strategies for Generative AI in Medical Communication Education: Exploring Effects on Role-Play Quality and Learner Interaction
0
Zitationen
4
Autoren
2025
Jahr
Abstract
<title>Abstract</title> <bold>Background</bold> : This study aimed to explore strategies for designing prompts that can support effective role-play in medical communication training using generative AI (ChatGPT) and to empirically analyze the impact of these strategies on interactions with AI. <bold>Methods</bold> : Medical communication role-play exercises based on challenging patient cases were conducted with 79 medical students, after which prompt-response data were collected through interactions with ChatGPT. A total of 142 prompts and AI responses were classified according to three criteria, namely, functional purpose, syntactic structure, and interaction outcomes. Subsequently, cross-tabulation and chi-square tests were conducted. <bold>Results</bold> : Prompts with the purpose of information provision elicited the highest success rate of responses from ChatGPT. Conditional syntax was used to induce effective responses for purposes such as emotion induction and role setting. The study found no statistically significant differences between interaction outcomes and dialogue success. Meanwhile, functional purpose and syntactic structure showed a significant association. In particular, interrogative syntax effectively responded to information provision, and conditional syntax, to emotion induction and role setting. <bold>Conclusions</bold> : The quality of GPT’s responses can vary strategically depending on the purpose and structure of the prompts. When using generative AI in medical communication training, prompt design strategies have a tangible impact on the quality of AI responses and success of interactions. In future educational settings, the effectiveness of AI-based training can be maximized through the elaborate design of prompt purposes and syntactic structures, which can contribute to improving the quality of immersive medical communication training.
Ähnliche Arbeiten
Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI
2019 · 8.231 Zit.
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
2019 · 8.084 Zit.
High-performance medicine: the convergence of human and artificial intelligence
2018 · 7.444 Zit.
Proceedings of the 19th International Joint Conference on Artificial Intelligence
2005 · 5.776 Zit.
Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)
2018 · 5.423 Zit.