Research Group Digital Communication and Transformation, 27170University of Duisburg-Essen, Duisburg, Germany.
Faculty of Business Administration and Economics, 9168Paderborn University, Paderborn, Germany.
Health Informatics J. 2021 Oct-Dec;27(4):14604582211052391. doi: 10.1177/14604582211052391.
The application of artificial intelligence (AI) not only yields in advantages for healthcare but raises several ethical questions. Extant research on ethical considerations of AI in digital health is quite sparse and a holistic overview is lacking. A systematic literature review searching across 853 peer-reviewed journals and conferences yielded in 50 relevant articles categorized in five major ethical principles: beneficence, non-maleficence, autonomy, justice, and explicability. The ethical landscape of AI in digital health is portrayed including a snapshot guiding future development. The status quo highlights potential areas with little empirical but required research. Less explored areas with remaining ethical questions are validated and guide scholars' efforts by outlining an overview of addressed ethical principles and intensity of studies including correlations. Practitioners understand novel questions AI raises eventually leading to properly regulated implementations and further comprehend that society is on its way from supporting technologies to autonomous decision-making systems.
人工智能（AI）的应用不仅为医疗保健带来了优势，也引发了一些伦理问题。现有的关于数字健康中人工智能伦理考量的研究相当匮乏，缺乏整体概述。一项系统的文献综述在 853 篇同行评议的期刊和会议上进行，共产生了 50 篇相关文章，这些文章被归类为五大伦理原则：有益、无害、自主、公正和可解释性。描绘了数字健康中人工智能的伦理格局，包括一个指导未来发展的快照。现状突出了潜在的领域，这些领域虽然没有经验，但需要研究。通过概述所涉及的伦理原则和研究的强度，包括相关性，验证了探讨较少的领域中仍然存在的伦理问题，并指导学者的努力。从业者最终会理解 AI 提出的新问题，从而导致对其进行适当监管的实施，并进一步认识到社会正从支持技术向自主决策系统发展。
