New publication on GRACE pilot study 2 among healthy older adults
Rasita Vinay, Ekaterina Uetova, Nora Camilla Tommila, Nikola Biller-Andorno, and Tobias Kowatsch have published a new study in JMIR Aging titled “A Hybrid Rule- and Large Language Model–Based Embodied Voice Assistant (GRACE) for Cognitive Stimulation in Older Adults: Usability Study Assessing Technical Feasibility, Technology Acceptance, and Working Alliance.”
In this pilot usability study, the team evaluated GRACE, an embodied voice assistant that integrates rule-based logic with large language model (LLM) capabilities to deliver structured cognitive stimulation to older adults. The study involved 21 healthy German-speaking adults aged 60 and over. Participants interacted with GRACE in a controlled laboratory setting and completed questionnaires and interviews to assess usability, perceived usefulness, enjoyment, and the working alliance between the user and the system.
The results show that participants rated GRACE positively across key user-experience dimensions, including ease of use, perceived usefulness, enjoyment, and intention to continue using the system — supporting the technical feasibility and strong user acceptance of hybrid conversational agents for cognitive health support. Areas for further improvement included personalization, response delays, and voice quality, pointing to important design considerations for future development.
This work contributes evidence toward the potential of hybrid AI-powered voice assistants to support cognitive stimulation and healthy aging, and sets the stage for further studies with populations at risk for dementia. This work is part of the larger GRACE study, funded by the Swiss National Science Foundation (Grant No. 10002915)