Designing Intimacy: Applying Norman’s Emotional Design Theory to Human-AI Companionship

AI companions, powered by advances in large language models and conversational AI, are increasingly designed for not only functional utility but also emotional engagement. As these systems evolve into entities capable of simulating intimacy, understanding the design mechanisms behind human-AI emotional relationships becomes critical. This study applies Norman’s three levels of emotional design to analyse how affective bonds with AI companions are formed, sustained, and deepened. By mapping each design level to stages of emotional interaction, the framework reveals psychological processes and ethical challenges, including authenticity, anthropomorphism, and projected intimacy. The analysis highlights the need for interdisciplinary approaches integrating design theory, psychology, and ethics, offering a structured basis for creating emotionally intelligent AI companions with both engagement and responsibility in mind.