Transforming dementia caregiver support with AI-powered social robotics

Area of Science:

• Gerontology
• Artificial Intelligence
• Human-Robot Interaction

Background:

• Informal dementia caregivers experience substantial emotional and physical strain.
• Existing evidence-based interventions, such as REACH, face limitations in scalability and high labor costs.
• There is a need for accessible and cost-effective support solutions for dementia caregivers.

Purpose of the Study:

• To design and evaluate a novel system, RISE, that leverages social robotics and generative AI to deliver automated and personalized training and stress management for dementia caregivers.
• To ensure the safety and accuracy of AI-generated content by grounding it in verified information (REACH Caregiver Notebook) using retrieval-augmented generative AI (RAG-AI).
• To assess the technical performance and user acceptance of the RISE system.

Main Methods:

• Development of the RISE system, integrating a social robot (Pepper) with a RAG-AI backend.
• The RAG-AI component was trained on the REACH Caregiver Notebook to provide safe and relevant information.
• The Pepper robot was utilized for interactive training modules, Q&A sessions, quizzes, and stress-reduction activities.
• Technical evaluation of the RAG-AI backend and a two-phase user evaluation of the RISE system were conducted.

Main Results:

• The RAG-AI backend demonstrated high performance, achieving 87% correctness and 92% relevancy against ground truth.
• User feedback indicated strong acceptance of the RISE system, with usability scores ranging from 3.6 to 4.6 out of 5.
• The system successfully delivered automated caregiver training and stress management interventions.

Conclusions:

• The integration of verifiable AI architectures with social robotics presents a feasible and scalable approach to improve dementia caregiver support and well-being.
• The RISE system shows promise as an effective tool for reducing caregiver burden and enhancing their quality of life.
• Future research should involve larger-scale studies with informal dementia caregivers to further validate the system's effectiveness.