US hospital giant turns to generative AI for clinical paperwork, raising safety and compliance questions

HCA Healthcare, one of the largest for‑profit hospital chains in the United States, is expanding its use of generative AI to draft clinical documentation such as visit notes, discharge summaries, and referral letters. The company has already been piloting systems that turn physicians’ voice notes into draft charts in selected emergency departments and inpatient units, and is now moving to broaden those pilots. The push is driven by persistent physician shortages and rising paperwork demands; executives hope AI will cut administrative time and free clinicians to see more patients.

The program, however, highlights familiar risks. Generative models can “hallucinate” incorrect information, and routing sensitive electronic health record data through external cloud services raises concerns about privacy and potential data breaches. HCA says that all AI‑generated drafts must be reviewed and edited by clinicians before being saved to the official record, and that only de‑identified data are used to further train models. Health‑policy experts remain cautious, questioning how robust human oversight will be in busy clinical environments and how forthcoming providers will be about error rates. With US regulators working on updated guidance for clinical AI tools, HCA’s rollout is likely to become a key test case for balancing efficiency gains with safety and accountability in hospital AI deployments.