This paper examines the working conditions and advantages of artificial intelligence (AI) and human intelligence (HI) concerning task input, processing, output, and constraints, leading to the identification of eleven AI-HI task characteristics. Based on these characteristics, we develop four AI-HI indices: AI Usability Index, AI Superiority Index, HI Superiority Index, and HI Safety Index. These indices inform a proposed typology of four task types: AI Non-Usable, AI Superior, AI-HI Collaborative, and HI Superior. This framework underpins a novel AI-HI task characteristic-classification model. Finally, taking accounting tasks as a case, the model is applied to analyze nineteen core accounting tasks, and assess their AI-HI characteristics, indices, and resultant classifications. The result can provide a rigorous analysis of the division of labor and substitution between AI and HI, offering a theoretical foundation for a transition from adversarial to symbiotic collaboration.

深入探讨人工智能(AI)与人类智能(HI)在任务输入、处理、输出以及制约因素等方面的工作条件和优势, 识别出11个AI-HI任务特征。在此基础上形成4个AI-HI指数——AI可用指数、AI优越指数、HI优越指数和HI安全指数, 并据此提出4种任务类型: AI不可用型、AI优越型、AI-HI协作型和HI优越型。通过上述工作, 构建一个新的AI-HI任务特征–分类模型。最后, 以会计工作为例, 应用此 AI-HI任务特征–分类模型, 逐一分析19个会计核心任务的AI-HI特征、AI-HI指数及其所属类型。研究结果为AI与HI的分工与替代问题提供科学分析, 为二者从对立走向协作共生提供理论依据。