Digital Support for Family Caregivers: Potential and Challenges of a Hypothetical AI Care Companion

Background/Objectives: Family caregivers play a central role in the provision of long-term home-based care and often provide unpaid support over extended periods. This role is associated with substantial psychological, physical, social, and financial burden. Despite high support needs, access to psychosocial services remains limited for many family caregivers. Against this background, AI-based care companions are discussed as a potential low-threshold supplement to existing support structures. The objective of this study was to explore subjectively perceived family caregiver burden and to examine expectations, acceptance conditions, and concerns regarding a hypothetical AI-based care companion, rather than to evaluate effectiveness. Methods: An exploratory mixed-methods study was conducted using an anonymous online survey. Perceived family caregiver stress was assessed using self-developed, non-validated ordinal items, including a single-item global burden rating and categorical stress domains. The questionnaire combined closed-ended items (Likert-scale and multiple-choice) with one open-ended question to assess perceived stress, experiences with psychosocial support, and attitudes toward a hypothetical AI care companion. Participants were recruited via an online caregiving course platform. Data collection was voluntary and anonymous and took place in Germany between October and November 2025. Quantitative data were analyzed descriptively and exploratorily, and qualitative responses were analyzed using thematic analysis. Results: Fifty-five family caregivers participated in the survey. Overall, perceived family caregiver burden was high, with psychological stress most frequently identified as the dominant stress domain. Difficulties in accessing psychosocial support were reported by 58% of the respondents. Willingness to consider using an AI-based care companion varied by degree of acceptance: 36% reported clear willingness, 31% expressed conditional or tentative willingness, and 33% indicated reluctance or rejection. The most frequently selected expected functions included emotional support, early detection of overload, and caregiving-related information. Data protection, professional reliability, and concerns regarding incorrect advice were identified as the most relevant perceived risks. Conclusions: The findings reflect family caregivers' perceived burden and anticipated needs, highlighting persistent gaps in psychosocial support. From the perspective of respondents, a hypothetical AI-based care companion could represent a complementary support option if it provides personalized, non-judgmental, and reliable assistance. These results describe perceived potential and acceptance conditions, not verified efficacy. Further research, including prototype development, usability testing, and pilot studies, is required to examine feasibility, ethical implications, and real-world impact.