Short answer: In a 2026 Pew Research survey of 3,488 US adults, about 73 percent accepted AI scheduling their appointments, while only 3 percent wanted AI diagnosing them alone and 41 percent rejected AI diagnoses entirely. Patients are not anti AI. They welcome it for the front desk and paperwork, and want a human making the medical decision. For a practice, that means put AI on the phones and the calendar, keep people on the care, and tell patients when AI is in the loop.
The headlines this week read like a problem for anyone using AI in a medical practice. "Americans wary of AI in healthcare," said CBS News, reporting on a fresh Pew Research Center survey. Six in ten people said they are uncomfortable with their doctor using AI for things like diagnosis and treatment. If you run a practice and you have been nervous about AI scaring patients off, that number probably confirmed your fear.
But read one layer down and the story flips. The same patients who do not want AI diagnosing them are, in large numbers, completely fine with AI booking their appointment. The survey did not find that people reject AI. It found that people reject AI in one specific place, the clinical decision, and welcome it almost everywhere else. For a practice owner, that is not a reason to back away. It is a precise instruction on where to put it.
What the survey actually found
Pew surveyed 3,488 US adults from June 22 to 28, 2026, and published the results in October. The questions were not "do you like AI," they were task by task, which is why the results are so useful. Here is the line patients drew.
| Healthcare task | Open to AI (alone or with human review) | Do not want AI involved |
|---|---|---|
| Scheduling appointments | About 73% (29% alone, 44% with review) | 18% |
| Prescription refills | Majority comfortable (around 64%) | Minority |
| Diagnoses | 50% (just 3% alone, 47% with review) | 41% |
| Insurance claim decisions | 43% (just 5% alone, 38% with review) | 46% |
Look at the top row against the bottom two. When AI is scheduling an appointment, four out of five people shrug and let it happen. When AI is diagnosing them, only 3 percent want it acting on its own, and the largest single group wants it nowhere near the decision. Insurance claims drew the most resistance of anything tested, with nearly half wanting AI to have no role at all, which makes sense given the headlines about denied claims.
That is not a contradiction. It is a very sensible line most of us would draw ourselves. I do not care whether a human or a bot finds me a Tuesday at 3pm. I care a great deal whether a human or a bot reads my scan. Your patients feel the same way, and now there is a large national survey proving it.
Why this is good news, not bad
Most of the AI a marketing agency would ever help a practice use lives entirely on the welcome side of that line. Answering the phone when the front desk is slammed. Booking and rescheduling. Texting a reminder so a patient does not forget Thursday. Catching the call that comes in at 9pm and would otherwise hit voicemail. None of that is diagnosis. None of it is an insurance decision. It is the administrative, behind the scenes work that the Pew data shows patients are not just tolerant of, but actively comfortable handing to a machine.
The fear that "patients will hate AI" comes from imagining the wrong use of it. Nobody is proposing a chatbot that tells a worried parent what is wrong with their kid. The realistic, valuable use is the boring stuff that your team is too busy to catch, and that patients are glad to have handled instantly instead of waiting on hold. We wrote before about how the front desk staffing gap is the real crisis in most practices, and this survey is the permission slip to fix it with AI without worrying you will spook people.
The one rule: keep AI off the clinical decision
The survey is just as clear about where not to go. Patients do not want AI making the medical call, and they especially do not want it quietly deciding anything about their coverage. That line matters for marketing too. Do not let your website, your intake, or your automations drift into anything that looks like AI giving medical advice or screening patients in or out. That is the one move the data says will cost you trust.
The healthiest setup is a clean handoff. AI handles the conversation up to the moment it becomes clinical, then a human takes over. A patient asks about availability, cost, parking, insurance accepted, what to bring, and AI answers all day and all night. The moment the question turns into "should I be worried about this symptom," a person steps in. That is the exact division of labor the American Medical Association has been pushing for too, with its recent policies insisting on human oversight and physician final say over AI, which we unpacked in what the AMA's AI guidance means for your practice.
Be transparent, because patients asked for it
A separate Pew analysis from August 2026 found Americans strongly want to be told when AI is used in their care. Some practices read that and panic, thinking they have to hide the AI receptionist. The opposite is true. Because your AI is on the welcome side of the line, telling people about it builds trust rather than breaking it.
"You can book anytime, even after we close, and our assistant will get you on the calendar" is a feature, not a confession. Patients like that the help is there at 9pm. What they dislike is discovering, after the fact, that a decision they thought a human made was actually a machine. Keep AI on scheduling and reminders, say so plainly, and transparency becomes a selling point. We made the same case in why being open about AI builds patient trust.
Where EtherealMinds fits
We work only with healthcare practices, and we build AI into exactly the places this survey says patients want it, and nowhere they do not. On the front end, our AI receptionist answers the calls your team misses, handles the questions, and books the appointment, then hands anything clinical straight to your people. Behind it sits the rest of the patient acquisition system: a fast website with online booking, reminders that cut no shows, and follow up that reactivates patients who went quiet. It is AI on the operations, humans on the medicine, which is the arrangement patients told Pew they are comfortable with.
The practices that win the next couple of years are not the ones that avoid AI, and they are not the ones that shove it into the exam room. They are the ones that read the room the way this survey just did: use it to answer the phone, fill the calendar, and free up your staff, and let your clinicians be the human patients still very much want to see.
Our honest take
The "Americans are wary of AI" headline is true and also a little misleading. People are wary of AI making their medical and money decisions, and they are right to be. They are not wary of AI making their life easier at the front desk. If you have been holding off on AI because you were afraid of the backlash, the survey just told you where the backlash actually lives, and it is not where you were planning to use it anyway. Put AI where patients welcome it, keep humans where patients need them, be honest about both, and you are not fighting the trend. You are riding the exact line the public drew.
Want AI in the right place in your practice?
Book a free strategy call. We will show you where AI can answer calls, fill your calendar, and cut no shows without ever touching the clinical side, and what it would look like in your practice. Healthcare only, no jargon, no pressure.
Book a free strategy call →