Intelligence amplification exploits the opportunities of artificial intelligence, which includes data analytic techniques and codified knowledge for increasing the intelligence of human decision makers. Intelligence amplification does not replace human decision makers but may help especially professionals in making complex decisions by well-designed human-AI system learning interactions (i.e., triple loop learning). To understand the adoption challenges of intelligence amplification systems, we analyse the adoption of clinical decision support systems (CDSS) as an organizational learning process by the case of a CDSS implementation for deciding on administering antibiotics to prematurely born babies. We identify user-oriented single and double loop learning processes, triple loop learning, and institutional deutero learning processes as organizational learning processes that must be realized for effective intelligence amplification adoption. We summarize these insights in a system dynamic model—containing knowledge stocks and their transformation processes—by which we analytically structure insights from the diverse studies of CDSS and intelligence amplification adoption and by which intelligence amplification projects are given an analytic theory for their design and management. From our case study, we find multiple challenges of deutero learning that influence the effectiveness of IA implementation learning as transforming tacit knowledge into explicit knowledge and explicit knowledge back to tacit knowledge. In a discussion of implications, we generate further research directions and discuss the generalization of our case findings to different organizations.

情报放大利用人工智能的机会，其中包括数据分析技术和编码知识，以提高人类决策者的智力。智能放大不会取代人类决策者，但可以通过精心设计的人机系统学习交互（即三环学习）帮助专业人士做出复杂的决策。为了理解智能放大系统的采用挑战，我们通过 CDSS 实施的案例分析了临床决策支持系统 (CDSS) 的采用作为组织学习过程，以决定对早产婴儿使用抗生素。我们确定面向用户的单循环和双循环学习过程、三循环学习、和机构deutero学习过程作为组织学习过程，必须实现有效的情报放大采用。我们在系统动态模型中总结了这些见解 - 包含知识库及其转换过程 - 通过该模型，我们分析了来自 CDSS 和情报放大采用的各种研究的见解，并通过该模型为情报放大项目的设计和管理提供了分析理论. 从我们的案例研究中，我们发现 deutero 学习的多重挑战会影响 IA 实施学习的有效性，因为将隐性知识转化为显性知识，并将显性知识转化为隐性知识。在讨论影响时，