Generative artificial intelligence in healthcare from the perspective of digital media: Applications, opportunities and challenges
Rui Xu,Zhong Wang
DOI: https://doi.org/10.1016/j.heliyon.2024.e32364
IF: 3.776
2024-06-05
Heliyon
Abstract:Introduction: The emergence and application of generative artificial intelligence/large language models (hereafter GenAI LLMs) have the potential for significant impact on the healthcare industry. However, there is currently a lack of systematic research on GenAI LLMs in healthcare based on reliable data. This article aims to conduct an exploratory study of the application of GenAI LLMs (i.e., ChatGPT) in healthcare from the perspective of digital media (i.e., online news), including the application scenarios, potential opportunities, and challenges. Methods: This research used thematic qualitative text analysis in five steps: firstly, developing main topical categories based on relevant articles; secondly, encoding the search keywords using these categories; thirdly, conducting searches for news articles via Google ; fourthly, encoding the sub-categories using the elaborate category system; and finally, conducting category-based analysis and presenting the results. Natural language processing techniques, including the TermRaider and AntConc tool, were applied in the aforementioned steps to assist in text qualitative analysis. Additionally, this study built a framework, using for analyzing the above three topics, from the perspective of five different stakeholders, including healthcare demanders and providers. Results: This study summarizes 26 applications (e.g., provide medical advice, provide diagnosis and triage recommendations, provide mental health support, etc.), 21 opportunities (e.g., make healthcare more accessible, reduce healthcare costs, improve patients care, etc.), and 17 challenges (e.g., generate inaccurate/misleading/wrong answers, raise privacy concerns, lack of transparency, etc.), and analyzes the reasons for the formation of these key items and the links between the three research topics. Conclusions: The application of GenAI LLMs in healthcare is primarily focused on transforming the way healthcare demanders access medical services (i.e., making it more intelligent, refined, and humane) and optimizing the processes through which healthcare providers offer medical services (i.e., simplifying, ensuring timeliness, and reducing errors). As the application becomes more widespread and deepens, GenAI LLMs is expected to have a revolutionary impact on traditional healthcare service models, but it also inevitably raises ethical and security concerns. Furthermore, GenAI LLMs applied in healthcare is still in the initial stage, which can be accelerated from a specific healthcare field (e.g., mental health) or a specific mechanism (e.g., GenAI LLMs' economic benefits allocation mechanism applied to healthcare) with empirical or clinical research.