The Brain Computer Interface (BCI) is a user interface that helps people with severe physical disabilities to communicate with others or to interact with their surroundings by using their brain waves to control devices. However, almost all conventional studies about stimulus-driven BCI focused on a visual modality. Until recently, there still haven't had useful tools to help blind people who suffer from severe physical disabilities (e.g., blind amyotrophic lateral sclerosis (ALS)) to communicate independently. Although there are some initial results, the accuracy of contemporary auditory stimulation modality BCI system is very poor. In this paper, the authors proposed an auditory stimulation modality using shadowing tasks to develop an auditory brain computer interface system (aBCI) with the accuracy above 70% to achieve the preliminary practicality. This aBCI system will be the foundation of the subsequent practical aBCI system.