Accuracy and Hallucination Rates of AI-Generated Clinical Notes From Home Health Conversations With Older Adults

Artificial intelligence (AI) ambient scribe systems use large language models to passively capture patient-clinician conversations and automatically generate draft clinical notes for clinician review. These tools are being adopted rapidly across US health systems and internationally; by mid-2025, nearly two-thirds of US hospitals using Epic electronic health records had implemented ambient AI documentation tools.1–3 In home health care (HHC), where older adults frequently present with multimorbidity, functional impairment, fall risk, polypharmacy, and caregiver dependence, the central question is whether these tools can safely document complex geriatric care without introducing false information or omitting critical facts.