Fewer Americans trust AI with their health care than did two years ago. More of them are using it anyway.
A new survey from Ohio State University's Wexner Medical Center found that only 42% of adults are open to AI being used as part of their care, down from 52% when the same survey ran in 2024.
Belief that AI can make health care more efficient dropped too, from 64% to 55%.
The survey, conducted by SSRS in mid-January among 1,007 adults, carries a margin of error of about 3.5 percentage points. That's a solid enough sample to treat the ten-point trust drop as a real shift, not statistical noise.
A Confidence Drop Nobody Expected
Dr. Ravi Tripathi, the medical center's chief health informatics officer, says the decline tracks a familiar pattern.
"When we first see something new and shiny, we think it's going to fix the world and replace health care and solve all of our medical problems," Tripathi said.
He expects the numbers to climb again over the next two to five years, as people settle into a clearer sense of where AI actually helps and where it doesn't.
The Gap Between Trust and Use
Here's where the survey gets genuinely strange. Despite the drop in stated trust, 51% of respondents said they'd used AI to make an important health decision without consulting a medical professional at all.
The specific use cases break down clearly. Sixty-two percent used AI to understand symptoms before deciding whether to seek care.
Forty-four percent used it to help explain a test result or diagnosis. A quarter used it to compare treatment options.
Dr. Hiten Patel, a family medicine physician at Wexner, thinks access is doing a lot of the work here. AI fills a gap when getting a doctor's appointment quickly isn't realistic.
He says it beats searching symptoms on Google or WebMD, since the information comes back more tailored to the person asking.
The Real Risk Isn't the Technology
Tripathi's bigger concern isn't that AI gets things wrong sometimes. It's what happens when a patient trusts it completely.
"We know that 2% of the time AI is going to be inaccurate or it will potentially hallucinate," Tripathi said.
"Physicians are not using AI 100%. We're not trusting it 100%. I would be really concerned about a patient who is following AI."
The artificial intelligence doesn't understand your story.
That's the gap no model closes. AI can summarize symptoms and explain lab results well.
It doesn't know a patient's history, family context, or the small inconsistencies a doctor might notice in person. That's the exact theme MAIN covered recently in a piece on what AI still can't replace in a physician's judgment.
Tripathi's recommendation is to treat AI as what he calls augmented intelligence: a tool that helps a patient show up to an appointment better informed, with the final medical decision still made alongside a doctor.
What This Means for Miami
South Florida's healthcare access gaps, driven by physician shortages in some specialties and a large population managing chronic conditions, make the appeal of AI symptom-checking obvious here too.
Miami's hospital systems and digital health startups building AI tools for patients should read this survey as validation with a warning attached. People want the convenience, using it heavily even as stated trust falls, but they're not distinguishing well between AI as a research aid and AI as a decision-maker.
Products and provider guidance that make that distinction explicit, rather than assuming patients will figure it out themselves, are likely to earn more durable trust than ones that don't.
For local health systems, that might mean something as simple as building a clear disclaimer into patient-facing AI tools: use this to prepare for your appointment, not to replace it. The survey suggests plenty of patients would welcome that guidance, even if they haven't been asking for it directly.