探討民眾採用聊天機器人作為輔助醫療工具之行為研究

A Study on the Factors Influencing Behavior of LINE Bot as an Assistive Healthcare Tool

黃維民(Wei-Min Huang)；吳旻蓁(Min-Zhen Wu)

聊天機器人 ； 醫療輔助 ； 統合分析模型 ； 信任 ； 感知智慧 ； LINE Bot ； Assistive Healthcare ； Meta-UTAUT ； Trust ； Perceived intelligence

電腦稽核

50期（2024 / 08 / 30）

52 - 66

繁體中文;英文

在公共衛生領域，人工智慧的最新進展使對話代理能夠與人類進行更真實的互動，聊天機器人等數位科技隱含的效率和規模水平為醫療保健提供了新的可能性，可以將個人層面的醫療照護擴展到人口層面。因此，未來有望為使用者提供有用的資訊和支援，從而促進民眾的福祉和實施更好的健康管理。本研究主要採用橫斷面問卷調查法進行資料蒐集，透過多種網路社群媒體平台，以Dwivedi et al. （2019）提出之統合分析模型（Meta - UTAUT）為基礎，加入外生變數「信任」、「感知智慧」進行實證研究。總共收集到263份問卷，剔除無效問卷後，有效回收率達到74.9%。統計方法是使用SmartPLS 4.0的PLSpredict程序進行預測和分析。依據檢驗結果顯示，13項假設當中共有9項成立，績效預期、社會影響對態度以及行為意圖的形成產生正面的影響。努力預期、便利條件、信任對行為意圖與便利條件對態度均不具顯著影響。民眾也會根據他們對醫療機構聊天機器人的態度來決定是否使用，這意味著，民眾對於工具的正面態度仍然會影響他們的實際使用行為。

In the field of public health, recent advances in artificial intelligence enable conversational agents to interact more realistically with humans, and the levels of efficiency and scale implicit in digital technologies such as LINE Bot offer new possibilities for health care that can transform care at the individual level extends to the population level. Therefore, it is expected to provide users with useful information and support in the future, thereby promoting people's well-being and implementing better health management. This study mainly uses a cross-sectional questionnaire survey to collect data through a variety of online social media platforms. It is based on the integrated analysis model (Meta-UTAUT) proposed by Dwivedi et al. (2019) and adds exogenous variables of trust and perceived intelligence. A total of 263 questionnaires were collected, and effective response rate reached 74.9%. The statistical method is to use the PLSpredict program of SmartPLS 4.0 for prediction and further analysis. According to the test results, a total of 9 of the 13 hypotheses were established. Performance expectations and social influence have a positive impact on the formation of attitudes and behavioral intentions. People will also decide whether to use LINE Bot based on their attitudes toward healthcare organizations which means that people's positive attitudes toward the tool will still affect their actual usage behavior.