A number of privacy breaches have occurred in recent years, which has made people pay increased attention to the security of information systems. On the basis of this issue, role-based access control (RBAC) has been proposed and proven through practice to be able to effectively guarantee the security of user system data. But, in RBAC, role engineering is a complex process. To simplify the process, an auxiliary interactive question-and-answer (Q and A) algorithm was proposed based on attribute exploration (machines and humans learn knowledge interactively). the auxiliary interactive Q and A algorithm based on attribute exploration has some defects. It is not only unable to work with many people, but also has difficulty finding qualified Q and A experts in actual work. To address these problems, this paper proposes an attribute exploration-based Role discovery model. This model not only avoids the time-consuming process in role engineering, but also solves the problem of the auxiliary interactive Q and A based on attribute exploration being unable to support multi-person collaborative question–answering. Therefore, the model algorithm can be used for machine learning knowledge to assist people to solve the problem of cross-departmental role formulation.