A new Pew Research Center survey of 3,488 U.S. adults finds that while a quarter of Americans have used AI chatbots to diagnose symptoms, most still want to know when their doctors use the technology. The findings point to a gap between growing adoption of generative AI in healthcare and public confidence in how it's applied.
The poll, conducted June 22-28, found that patients primarily turn to chatbots for convenience, affordability, and speed. Many use them to quickly interpret lab results, better understand a doctor's diagnosis, or decide whether to seek in-person care. Among those who have tried it, 47% said the health information generated was "extremely" helpful, and 48% said it was "somewhat" helpful. Only 5% said the results were not helpful at all.
Convenience vs. accuracy
Chatbots work by simulating a conversation and responding to prompts using the data they were trained on. That allows 24/7 access at no cost, often without an account. But the same technology is known to "hallucinate" - generating false information with high confidence. For medical questions, a confident wrong answer can be dangerous.
That trade-off is why patient trust remains limited, even as usage grows. About one in three U.S. adults have used chatbots for health-related questions, according to recent numbers cited in the survey. ChatGPT and similar tools are increasingly part of everyday health decisions, but the Pew findings suggest Americans view them as a useful supplement, not a replacement for professional care.
Providers are using AI too
Beyond patients asking chatbots questions, medical providers are using AI to take notes, interpret exams, and triage patients. The use is already delivering faster results in some settings: AI has been used to cut X-ray interpretation turnaround from 11.2 days to 2.7 days.
Patients want all of that disclosed. According to the survey, 72% said it is "extremely or very important" for their medical provider to disclose any use of AI. Eight in 10 people expect same if AI is being used for diagnoses or analysis of scans, X-rays, and lab results.
Yet nearly half the respondents - 47% - said they are unsure whether AI has ever been used in their care. This gap between expectation and awareness matters for AI for Healthcare practitioners, who will need to build transparency into workflows rather than treating it as an afterthought.
Mixed comfort with sharing health information
Public comfort with the technology remains mixed. The survey found that 29% of Americans say they are extremely or very comfortable sharing health information with AI chatbots, while 26% say they are uncomfortable. The rest are somewhere in the middle.
These results are not surprising. AI chatbots are trained to produce a probable answer, not necessarily one grounded in truth. So patients are showing a willingness to use them for general questions and basic guidance, but they want to know when AI is making decisions that affect their medical care.
Why this matters for healthcare professionals
For doctors, nurses, and clinic administrators, the takeaway is straightforward: patients expect to be told when AI is used in their care. Transparency isn't a luxury at this point - it's a prerequisite. Anyone integrating AI into their workflow should treat disclosure as a standard part of patient communication, like explaining a diagnosis or a treatment plan. Clinics that make their AI policies visible will be better positioned to build then trust that the data shows is still missing.
Your membership also unlocks:
