In this book chapter, we discuss the human-centered perspective on the design of intelligent systems and argue that intelligent systems can be designed and evaluated in the same manner as any other interactive system. We must identify the right stakeholders and their core tasks and clearly identify the user needs to derive the correct requirements for such intelligent systems. We introduce the concept of human-centered quality and continue with two essential concepts: Context of Use: This is the context in which an intelligent system is used. A context of use description identifies user profiles and core tasks to define requirements for specific user groups and their tasks. As we will see, intelligent systems can take over some of these tasks, but they also may introduce new (meta) tasks that must be supported by the user interface of the intelligent system. Stakeholder Requirements: In modern requirements engineering, human-centered requirements are derived from stakeholder needs. The portion of the stakeholders who use the intelligent system (i.e., the users) are success-critical in terms of acceptance and productivity for the organization. However, there are also indirect users and other organizational roles such as business owners, people managers, and team supervisors. This stakeholder view is important when discussing ethical issues such as intelligent systems’ purpose, fairness, authority, and controllability.