Patients Demand Provider Oversight as AI in Healthcare Expands: 56% Accept AI Diagnosis With Human Control

Patient Comfort Hinges on Provider Involvement, Not AI Autonomy

Patient comfort with AI in healthcare depends entirely on provider oversight, according to new data from the West Health-Gallup Center on Healthcare in America. The survey of over 5,000 U.S. adults found that 56% are comfortable with AI making diagnosis or treatment decisions when their healthcare provider can change or override them, but that figure collapses to just 9% when AI makes those decisions without provider review1

1

. This pattern holds across clinical care applications, with patient comfort dropping from 53% to 13% for imaging review and from 49% to 11% for medication decisions when provider involvement is removed2

2

.

The findings reveal that AI-driven health care faces a fundamental trust barrier. Patients want their providers engaged, not replaced. When AI assists and the provider makes the final decision, comfort levels range from 38% to 52% depending on the clinical application1

1

. "AI has enormous potential to improve health care, but patients are sending a clear message that they are not afraid of using the technology and want their providers involved, not replaced," said Timothy Lash, president and chair of the West Health Policy Center1

1

.

AI Adoption Outpaces Training as Communication Gap Emerges

While AI adoption in healthcare accelerates, provider training lags dangerously behind. A separate survey of more than 1,000 healthcare providers found that nearly nine in 10 (89%) report using AI for work in the past year, yet only 20% have received any formal AI training1

1

. Training rates vary dramatically by practice setting, from 30% in hospitals or health system practices to just 11% in independent practices1

1

.

Providers primarily use AI for gathering information about medications (78%) or symptoms and conditions (76%). More than half (53%) have used it to review patient-specific information like test results or imaging, while 50% use it to write clinical notes1

1

. Nearly half use AI to prepare patient instructions or after-visit summaries (49%), summarize medical records (47%), or record and transcribe appointments (45%)1

1

.

A significant communication gap exists between providers and patients regarding AI acceptance. Providers underestimate patient comfort with AI decisions by 20 percentage points. While 56% of patients are comfortable with AI making diagnosis and treatment decisions that a provider can override, only 36% of providers believe most of their patients would be comfortable with this arrangement1

1

.

Even AI-Savvy Patients Demand Human Oversight

The preference for provider involvement extends even among patients most receptive to AI. Among the 39% of patients who have used AI to gather physical or mental health information or advice in the past year, over half are comfortable with AI assisting their provider, but only 12% to 16% are comfortable with clinical advice from AI without that assistance1

1

. Similarly, among the 23% of patients who trust AI-generated health information, 65% to 71% say they are comfortable with its use in clinical care when a provider provides oversight, compared with just 18% to 20% who are comfortable without provider review1

1

.

When asked what would make them more comfortable with AI making autonomous decisions, fewer than half of patients selected the ability to request provider review as helpful. No other factor tested—including strong privacy protections, evidence of AI's safety and effectiveness, cost savings, or faster decisions—would make more than a third of uncomfortable patients more comfortable2

2

. This suggests that provider oversight is not merely one preference among many, but a fundamental requirement for patient trust in AI.

"These findings suggest reason for optimism about the role of AI in health care," said Joe Daly, global managing partner at Gallup. "The data show that patients are open to AI having a role in their care, particularly when they know a clinician can still step in. As AI becomes more integrated into health care, understanding where patients are comfortable with its use versus where they still value human oversight will be critical"1

1

. The research makes clear that successful AI integration requires closing the training gap and maintaining provider involvement in AI decisions.