Emotion-AI, also known as affective computing, has evolved from a niche scientific field into a core capability for modern interactive information systems. By interpreting emotional cues through speech, facial expressions, gestures, biosignals, and contextual behavior, Emotion-AI enables systems to adapt interactions based on users' affective states. This paper examines the theoretical foundations, technological architecture, challenges, and applications of Emotion-AI integration in interactive information systems. A conceptual framework is proposed highlighting four layers of integration—affective sensing, affect modeling, adaptive interface behavior, and ethical governance. Through examples across healthcare, customer service, education, entertainment, and workplace analytics, the research demonstrates how Emotion-AI can enhance personalization, engagement, empathy, and decision support. The paper also addresses ethical concerns such as emotional manipulation, bias, privacy, and consent.