This chapter provides a comprehensive exploration of integrating ethics into AI system architecture using the AI Reference Architecture framework. It emphasizes embedding ethical principles—such as transparency, privacy, fairness, accountability, and human-centeredness—directly into the design phase of AI systems. Key approaches include creating ethical guidelines, designing reference architectures, and implementing practical strategies like bias mitigation, explainability, and privacy-preserving techniques. Through case studies and practical applications, such as chatbots, this chapter demonstrates how ethical principles can be operationalized in real-world AI systems. The chapter also highlights challenges, such as balancing ethical constraints with performance, and offers future directions for advancing ethical AI design.