Establishing a long-term human presence on the Moon and beyond remains a central goal of space exploration, with programs like Artemis paving the way for Mars missions. These ambitious endeavors face challenges, including communication latency, bandwidth constraints, and autonomous operations in safety-critical environments. Future astronauts and mission control will rely heavily on artificial intelligence (AI) systems to process and integrate large, fragmented datasets from spacecraft, rovers, and life support systems, making human-centered AI (HCAI) essential to safeguard alignment with human goals. This chapter explores the concept of controllability, elaborating on the importance of ensuring that human users retain the ability to understand, direct, and intervene in the behavior of AI systems. Two key HCAI paradigms that support this goal are examined: human-in-the-loop (HITL) and human-on-the-loop (HOTL). HITL systems prioritize continuous human engagement to adapt AI-driven processes in real time, while HOTL systems focus on autonomous operations with periodic human oversight. Both paradigms maintain human accountability and agency, offering complementary approaches to managing dynamic, safety-critical tasks during long-duration spaceflight missions. Taking a holistic approach, the chapter explores both backend and frontend components of HITL and HOTL systems. The backend includes algorithms, data handling, and supervisory control mechanisms that underpin AI functionality. The frontend encompasses user interfaces, augmented reality (AR), virtual reality (VR), and explainable AI elements that enable intuitive and transparent human-AI collaboration. Case studies, including the Checklist Organizer for Research and Exploration (CORE) personal assistant for astronauts developed at the European Space Agency (ESA) and interfaces for spacecraft science operations at NASA, illustrate practical applications of these principles. By addressing both theoretical and applied aspects, this chapter closes with challenges, future research topics, and guidelines for designing HITL and HOTL systems that ensure robust human-AI collaboration in future human spaceflight, supporting human oversight and adaptability.