What the study found
The study found that emotion AI, or artificial intelligence that adapts to users' emotions, behaviors, and context, can be explored as part of eco-feedback in home personal assistants. The co-design sessions produced eight design ideas, including emotion-adaptive framing, emotion-timed interaction and delivery, and emotion-aware environment and social adaptation.
Why the authors say this matters
The authors say this matters because they see limited research on how people perceive emotion-adaptive eco-feedback and on how emotion AI might be adopted in real-world home settings. The study suggests that the ideas generated in co-design can offer new insights into both emotion AI for eco-feedback and AI co-design methods.
What the researchers tested
The researchers used a co-design method called Matchmaking for AI, which brought together real users and researchers. They created a living lab with 11 participants in Germany over half a year, ran a pre-interview about users' behaviors, requirements, and expectations for eco-feedback, and later held a co-design session based on appliance energy consumption data collected through smart plugs and the open.DASH platform.
What worked and what didn't
The co-design process generated eight concepts for integrating emotion AI into eco-feedback. These included emotion-adaptive eco-feedback framing, emotion-timed interaction and delivery, and emotion-aware environment and social adaption. The abstract does not report comparative testing of these ideas or say which worked better than others.
What to keep in mind
The available summary does not describe limitations in detail beyond noting that research in this area is limited. The abstract also does not report whether the design ideas were implemented, tested in use, or evaluated for outcomes such as well-being or energy efficiency.
Key points
- Emotion AI was explored for use in home eco-feedback within personal assistants.
- A co-design method called Matchmaking for AI was used with 11 participants in Germany.
- The study produced eight design ideas for emotion-adaptive eco-feedback.
- The ideas included emotion-adaptive framing, timed interaction and delivery, and environment and social adaptation.
- The abstract does not report direct testing or comparison of the design ideas.
Disclosure
- Research title:
- Emotion-adaptive eco-feedback concepts were generated for home assistants
- Authors:
- Lu Jin, Apostolos Vavouris, Nico Castelli, Dominik Pins, Alexander Boden, Lina Stankovic, Vladimir Stanković
- Institutions:
- Fraunhofer Institute for Applied Information Technology, University of Strathclyde
- Publication date:
- 2026-03-16
- DOI:
- 10.1145/3789680
- OpenAlex record:
- View
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.