Through coexistence with Self-Mirroring Twins (SeMT), this project establishes the science and technology that help people reflect on themselves and change their behavior autonomously.
Background and Objectives
When AI and robots merely replace human actions, they may weaken people's autonomy to think, choose, and act for themselves. Rather than directing users unilaterally, this project supports lifestyle improvement and rehabilitation through interaction with “another self” that sometimes encourages them and sometimes offers a different perspective.
Project Structure
Virtual SeMT
A digital “other self” in VR, AR, or on a smartphone supports awareness and behavior change in everyday life.
Related: Research Area 1
Physical SeMT
Systems that act on the body, including rehabilitation robots and prosthetic devices, support autonomous movement.
Related: Research Area 2
Cross-Cutting Scientific and Technical Foundations
Behavior-generation models that reflect users’ internal states, together with methods for evaluating empathy and autonomy, are applied across both the Virtual and Physical domains.Related: Research Areas 3 and 4
Social Acceptance and Institutional Design
We study ethical issues, public dialogue, technology-design guidelines, and governance to enable responsible implementation in society.Related: Research Area 5
Five Research Areas
Research Area 1 | Platform Development for Virtual SeMT and Social Behavior Change
Lead: Tetsunari Inamura
We develop the SeMT system platform using VR, AR, and smartphones, along with models for measuring users’ everyday behavior and generating behavior. In the latter half of the project, we will conduct validation studies of behavior change through social interaction with multiple SeMTs on smartphones.
Research Area 2 | Applying SeMT to Physical Assistive Systems and Field Validation
Lead: Shingo Shimoda
Using rehabilitation robots and prosthetic devices, we investigate the relationship between body ownership, motivation, self-efficacy, and Physical SeMT behavior strategies. We aim to improve motor performance without compromising user autonomy and conduct field validation in care facilities.
Research Area 3 | Adaptive and Empathetic SeMT Behavior Generation Engine
Lead: Takato Horii
We construct computational models that reflect human internal states and evaluate behavior-change mechanisms that account for emotion. Based on the free-energy principle, we establish SeMT behavior-generation algorithms that enhance self-efficacy.
Research Area 4 | Evaluation Framework for Empathy toward SeMT and User Autonomy
Lead: Tetsunari Inamura
Using brain activity and physiological signals, we estimate users’ internal states and establish objective and subjective evaluation methods for adapting SeMT behavior in real time. We calibrate individual differences in the boundary between self and other, linking the results to designs that balance empathy and autonomy.
Research Area 5 | Social Acceptance and Institutional Design for SeMT
Leads: Makoto Kureha and Minoru Asada
We identify ethical issues associated with SeMT and possible responses, and evaluate social acceptance through public dialogue. We propose guidelines for ethically and socially acceptable technology design and support implementation through agile governance.
Areas for Collaboration
- Human–AI / Human–Robot Interaction
- Digital twins, XR, and personal AI
- Rehabilitation, movement assistance, and prosthetic devices
- Health behavior and lifestyle improvement
- Behavioral science, cognitive science, ethics, and social acceptance
Toward Collaborative Research and Social Implementation
We seek collaborative research with universities, research institutions, and companies by bringing together validation sites, data, devices, AI technologies, and evaluation methods.