AI acceptance and trust among family caregivers as surrogate decision-makers

Purpose This study aims to explore the joint influence of all four factors (performance expectancy, perceived risk, facilitating conditions and trust in AI), along with the remaining two factors (behavioral intention and caregiver health literacy), on caregiver AI acceptance. Design/methodology/approach To test both net-effect paths and configurational pathways, a cross-sectional survey of family caregivers was analyzed using partial least squares structural equation modeling and fuzzy-set qualitative comparative analysis (fsQCA) was carried out. Findings The findings revealed that all the hypothesized direct relationships were significant (p < 0.05), with the model accounting for 55.6% of the variance in behavioral intention, 46.0% in Trust in AI and 39.6% in actual adoption behavior. The number of all possible configurations (solution coverage = 0.965) identified by fsQCA was five. Research limitations/implications Cross-sectional, nonprobability design limits causal inferences and generalizability of the study. Practical implications Results drive clear risk communication design, progressive caregiver induction according to adoption type, literacy-aware institutional assistance and equipped uncertainty revealed in future AI nursing equipment. Originality/value The study’s reconceptualization of the perceived risk as an integrative signal and the inclusion of variance-based structural equation modeling and configurational analysis reveal the substitutive logic used to reach the same adoption outcome for differing caregiver subgroups.