University of California, Irvine, Irvine, CA, USA.
AMIA Annu Symp Proc. 2021 Jan 25;2020:1170-1179. eCollection 2020.
Previous research has studied medical professionals' perception of artificial intelligence (AI). However, there has been a limited understanding of how healthcare consumers perceive and use AI-powered technologies such as mobile health apps. We collected 40 popular mobile health apps that claim to have adopted AI, to study how AI is explained in these apps' descriptions, and how users react to it through app reviews. We found that four AI features (Recommendation, Conversational Agent, Recognition, and Prediction) are frequently used across seven health domains, including Fitness, Mental Health, Meditation and Sleep, Nutrition and Diet, etc. Our results show that (1) users have unique expectations toward each AI features, such as including feedback for recommendations, humanlike experience for conversational agents, and accuracy for recognition and prediction; (2) when AI is not adequately described, users make their own attempts to understand AI and to find out how (well) it works.
先前的研究已经研究了医学专业人士对人工智能（AI）的看法。然而，对于医疗保健消费者如何感知和使用 AI 驱动的技术（如移动健康应用程序），人们的了解有限。我们收集了 40 个流行的声称采用 AI 的移动健康应用程序，以研究 AI 在这些应用程序描述中是如何解释的，以及用户如何通过应用程序评论对此做出反应。我们发现，四个 AI 功能（推荐、对话代理、识别和预测）在七个健康领域中经常使用，包括健身、心理健康、冥想和睡眠、营养和饮食等。我们的研究结果表明：（1）用户对每个 AI 功能都有独特的期望，例如推荐功能需要包含反馈，对话代理需要有类似人类的体验，识别和预测功能需要准确性；（2）当 AI 没有被充分描述时，用户会自己尝试理解 AI 并了解它是如何工作的。
