Experts stress keeping a "human in the loop" as AI use grows in U.S. healthcare
A majority of U.S. physicians now use AI, but public trust remains low; experts advise patients to ask specific questions to ensure human oversight of AI tools in their care.
A recent KFF poll showed only 8% of Americans have a great deal of trust in AI managing appointments or health records, while 32% would trust an online AI health tool. In contrast, a 2026 survey found more than 80% of U.S. doctors use AI professionally, double the share from 2023. Clinicians cite benefits such as more accurate notes, a second opinion for diagnoses, and continuous monitoring of at-risk patients.
The gap between patient confidence and clinician enthusiasm has prompted calls for a “human in the loop” approach, where AI assists but does not replace medical judgment. Researchers and health-system leaders say patients feel more comfortable when doctors use AI as an aid rather than an autonomous decision-maker.
"Doctors… and nurses and staff have always been interested in making the best decision for the people under their care — and these tools can help with that," said Alison Callahan, a research scientist at Stanford University.
"AI is new, but the trust that serves as the foundation of the physician-patient relationship is not," added Timothy Keyes, a machine-learning scientist at Stanford Health Care.
AI applications range from large-language-model chatbots that patients consult about symptoms to hospital systems that use AI to summarize visit notes, schedule follow-ups, or predict complications. Jennifer Goldsack, CEO of the Digital Medicine Society, noted that adoption remains highly localized, depending on individual doctors or health systems.
Experts recommend that patients ask four key questions: whether AI will be used in their care and how; how their data will be protected; whether they can opt out; and how the system ensures the AI works as intended. Transparency from both sides—physicians explaining AI use and patients disclosing any prior AI-based self-diagnosis—can reinforce the collaborative “human in the loop” model.
The overall message is that open, honest conversations about AI use are essential for maintaining trust and ensuring that technology enhances, rather than replaces, clinical decision-making.