Fifty-six percent of U.S. adults who visited a primary care provider in the past year are comfortable with artificial intelligence making medical decisions, provided their doctor can override the system.
This finding, published in a November 2026 Gallup report, reveals a significant gap between patient willingness to use AI and the perceptions of primary care physicians.
The survey, conducted from August 12 to 24, 2026, via Gallup News Panel with 5,330 U.S. adults, carries a margin of sampling error of plus or minus 1.7 percentage points. While 56% of patients approve of AI-led diagnosis and treatment decisions when a provider holds final authority, only 9% are comfortable with AI making such decisions without any review.
For other clinical tasks, the preference for oversight remains consistent:
- Medication decisions: 49% of patients are comfortable with AI if a provider can make the final call, compared to 11% without oversight.
- Medical imaging review: 53% are comfortable with AI assistance under provider supervision, while 13% accept autonomous AI review.
The research, conducted in partnership with West Health through the West Health-Gallup Center on Healthcare in America, also tested what factors might increase comfort for those hesitant about AI.
When asked about potential factors like strong privacy protections, evidence of effectiveness, lower costs, or faster decisions, none would make more than a third of those patients more comfortable. By contrast, the data shows that 38% of patients are comfortable when AI assists and the provider makes the final decision.
Despite high patient receptivity, primary care physicians often underestimate how comfortable their patients are with AI-assisted care.
For diagnosis and treatment, providers estimated that only 36% of patients would be comfortable with AI-led decisions that they could override—a gap of 20 percentage points compared to the actual patient response of 56%. Similar discrepancies appeared regarding medication decisions, where providers estimated 42% comfort compared to the 49% reported by patients, and imaging review, where providers underestimated comfort by 11 points.
“These findings highlight the growing role of AI-enabled clinical information platforms in clinical practice and suggest that efforts to evaluate, teach, and integrate AI in clinical medicine should be informed by empirical studies of its use in actual practice,” the researchers said in the study.