Health information seeking with generative AI among adults with diabetes or prediabetes

Background Generative artificial intelligence (GenAI) may provide accessible health information for adults with diabetes or prediabetes, but its use for diabetes-specific needs remains insufficiently understood. Objective To examine diabetes-specific GenAI information seeking and assess whether its association with GenAI health-information use is statistically accounted for by perceived utility, critical literacy, verification behavior, behavioral impact, and trust/risk awareness. Methods A cross-sectional online survey was conducted between May 5 and May 20, 2026, among 617 Israeli adults with diabetes or prediabetes recruited through a research panel. Participants completed questionnaires assessing general and diabetes-specific GenAI health-information seeking, five experiential dimensions of GenAI use, and background variables. Pearson correlations and a parallel mediation/path model were examined, adjusting for age, education, diabetes type, technology literacy, non-AI online health-information seeking, and trust in GenAI medical advice. Results Diabetes-specific GenAI information seeking was low (mean=1.82, SD=1.02) and positively associated with GenAI health-information use (r=.547, p<.001). In the adjusted model, perceived utility (β=.134, p=.032) and behavioral impact (β=.201, p<.001) were positively associated with diabetes-specific information seeking, whereas trust/risk awareness was negatively associated (β=−.132, p<.001). GenAI use remained directly associated with diabetes-specific information seeking (β=.324, p<.001). The model explained 40.2% of the variance. Conclusions GenAI may serve as a supplementary information resource for adults with diabetes or prediabetes. Diabetes education should address the critical and safe use of GenAI-generated health information.