A June survey of nearly 5,000 U.S. adults by the Pew Research Center found that more than 70% of participants believe it is extremely important for healthcare providers to disclose when artificial intelligence is used in diagnostic decisions. The findings arrive as health systems nationwide rapidly adopt AI tools for everything from analyzing medical scans to transcribing patient conversations, often without clear patient communication about their use.
"Trust, above all, shapes Americans' views on AI in healthcare," said Andrew Crawford, senior counsel of data and privacy at the Center for Democracy and Technology. He emphasized that patients need to understand how their data is protected and trust that it won't be sold to data brokers or used for advertising.
The demand for transparency extends beyond clinical decisions. Most survey participants said they want to be told when AI handles administrative tasks such as taking notes during appointments, ordering prescription refills, or scheduling visits. Nearly half of those surveyed said they are not sure if AI is used in hospitals or doctors' offices at all.
The gap between deployment and disclosure
Patient information entered into AI systems - whether an AI scribe or a diagnostic tool - falls under the Health Insurance Portability and Accountability Act privacy rule. But Crawford said knowing what tools providers deploy is essential for patients to make informed choices. "People should have more confidence about when and where their healthcare providers are deploying AI tools," he said. Patients can then ask what role AI plays, how it works, what role the doctor retains, and what alternatives exist if they choose to opt out.
An ongoing class action lawsuit against Sutter Health and Memorial Care has brought these tensions into sharp relief. Plaintiffs alleged that medical staff used AI to record confidential physician-patient communications without asking permission, in violation of state and federal law. The tool at the center of the lawsuits is Abridge AI, a real-time clinical documentation platform adopted by numerous health systems. Kaiser Permanente deployed it across 40 hospitals and more than 600 medical offices in 2024, and Mayo Clinic rolled it out to roughly 2,000 clinicians serving over 1 million patients in early 2025. Kaiser Permanente said the tool requires patient consent, and AI-generated notes are reviewed by a human before entering medical records.
Accuracy concerns in clinical and administrative tools
Crawford pointed to accuracy as another patient concern. An AI scribe can help doctors focus more on the conversation, but transcription errors could lead to incorrect diagnoses or prescriptions. This is especially relevant for professionals managing AI for Medical Records Clerks, where accuracy in documentation directly affects patient safety and downstream clinical decisions.
"Doctors deploying these tools, even the administrative tools, have to make sure that the tools are fit for purpose, are accurate, and that they are testing them and evaluating them not only before they deploy them, but once they're deployed as well, to make sure that they're functioning as the doctor wants them to," Crawford said.
Why this matters for healthcare professionals
The survey data and legal challenges signal that disclosure is becoming a baseline expectation, not a differentiator. For clinicians, practice managers, and health IT staff, the immediate takeaway is practical: review how your organization informs patients about AI use in both clinical and administrative workflows. Consent processes, signage, and verbal explanations during intake all face scrutiny. The liability risk is not theoretical - it is already playing out in court. Training investments in AI for Healthcare can help teams evaluate tools for accuracy, establish disclosure protocols, and build the patient trust that the Pew data shows is currently fragile.
Your membership also unlocks:
