Reimagining Emotion AI at Home: Exploring the Potential of Emotion-adaptive Eco-feedback in Personal Assistant Using Matchmaking for AI

Emotion Artificial Intelligence (AI) is transforming the capabilities of personal assistant by enabling real-time adaptation to user emotions, behaviours, and contextual needs. This paper explores the potential of emotion-adaptive eco-feedback in personal assistant, particularly within home environments, to foster well-being, energy efficiency, and personalised user experiences. Currently, there is limited research on how users perceive emotion-adaptive eco-feedback and how emotion AI can be adopted in the eco-feedback within personal assistant in real-world settings. To address this, we employed a co-design method — Matchmaking for AI — to facilitate collaboration between real users and researchers. We built a living lab with 11 participants in Germany for half a year and conducted two experimental sessions: a pre-interview to understand user behaviours, requirements, and expectations on eco-feedback, and a co-design session using matchmaking for AI after half a year based on their appliance energy consumption data collected by our smart plugs using our open. DASH platform. The co-design sessions collaboratively brainstorm ideas for potential emotion AI adoptions and identify what their needs should be addressed by emotion AI technology. Through a co-design session, we generated eight design ideas that integrate emotion AI into eco-feedback. These concepts include emotion-adaptive eco-feedback framing, emotion-timed interaction and delivery and emotion-aware environment and social adaption. Our work explores the potential of using Emotion AI in eco-feedback within personal assistant and also provides new insights into AI co-design methodologies.