Abstract: The application of artificial intelligence technology in the field of accounting is in the exploratory stage of human-machine hybrid collaboration. The dominance and different roles played by humans and intelligent machines in work tasks have formed diverse human-machine collaboration modes. It is necessary to explore in depth which human-machine collaboration mode to apply in accounting tasks of different complexities. This study investigates the impact of accounting task complexity on the choice of human-machine collaboration mode by designing a mixed experimental scenario of 2 (task complexity)×2(human-machine collaboration mode). It is found that when faced with complex accounting tasks that require deep reasoning or processing with human thought, people tend to have a higher sense of credibility in the accounting processing results of the human dominated human-machine collaboration mode, and thus are more inclined to choose this collaboration mode; When faced with simple accounting tasks that can be regularized, people have a higher awareness of the credibility of accounting processing results under the human-machine collaboration mode dominated by artificial intelligence, resulting in a higher willingness to choose this mode. The research conclusion provides inspiration and reference for how to choose appropriate human-machine collaboration modes according to task needs in enterprise accounting work, and reveals the mechanism of trust perception.

Key words: Complexity of Accounting Tasks, Human-machine Collaboration Mode, Willingness to Choose, Perceived Credibility

摘要： 会计领域对人工智能技术的应用处于人机混合协作的探索阶段,人与智能机器在工作任务中的主导权及扮演角色的不同,形成了多样的人机协作模式,不同复杂性的会计任务中应用何种人机协作模式是需要深入探究的问题。本文通过设计2(任务复杂性)×2(人机协作模式)的混合实验情景来考察会计任务复杂性对人机协作模式选择的影响。研究发现,当面临需要用人类的思想进行深入推理或处理的复杂会计任务时,人们会对人类主导的人机协作模式所作会计处理结果形成更高的可信性感知,从而更倾向于选择该协作模式;当面临可规则化的简单会计任务时,人们则对人工智能主导的人机协作模式下的会计处理结果的可信性感知更高,从而产生更高的选择意愿。研究结论为企业会计工作中如何根据任务需要选择合适的人机协作模式提供启发和参考,并揭示了信任感知的作用机制。

关键词: 会计任务复杂性, 人机协作模式, 选择意愿, 可信性感知