Americans Want Transparency and More Say When AI Is Used in Their Healthcare
核心洞察
A Pew Research Center (搜索) survey of 3,488 U.S. adults found 72% say it is extremely or very important that providers tell them when AI is used in their healthcare.
Eight-in-ten or more want to be informed when AI is used for tasks directly affecting care, such as analyzing scans, diagnosing conditions, or explaining lab results.
A majority (53%) feel they have little to no say over AI use in their care, and Americans are three times as likely to want more input than to be comfortable with current levels (63% vs. 21%).
A new Pew Research Center (搜索) survey of 3,488 U.S. adults, conducted June 22–28, 2026, finds that Americans are broadly uncertain about whether artificial intelligence is being used in their healthcare—and they overwhelmingly want to be informed when it is. About seven-in-ten U.S. adults (72%) say it is extremely or very important that a doctor or other healthcare provider tells them if they are using AI in their healthcare, with majorities across age, gender, education, and racial and ethnic groups expressing this view.
The findings arrive as discussion rapidly evolves about medical professionals' legal and ethical obligations to obtain patients' consent before using AI tools.
Strong Demand for Disclosure, Even for Administrative Tasks
The desire for transparency is strongest when AI directly shapes medical decisions. Eight-in-ten or more Americans want a provider or their office to tell them if AI is used for tasks that can directly affect care, such as analyzing medical scans, diagnosing conditions, and explaining lab results.
When AI is used to take notes in the background during an appointment, 72% of Americans say they should be told. Smaller shares—but still more than half—feel this way about behind-the-scenes administrative activities, including scheduling medical appointments (56%) and getting prescription refills (64%). Scheduling is the task that the largest share feels they do not need to be told about (33%). For each task surveyed, another 9% to 10% say they are not sure whether they need to be informed about AI use.
White adults were more likely than Black or Hispanic adults to say they want to be told if AI is used for the three tasks that directly shape medical decisions.
Limited Sense of Control Over AI in Care
Americans tend to feel they lack control over AI use in their healthcare. The largest share (53%) say they have not too much or no say about whether a provider uses AI in their care, while another 33% feel they have at least some say. White adults (60%) are more likely than Hispanic (40%), Black (43%), and Asian adults (42%) to say they have little to no say.
Americans are three times as likely to say they want more input on whether AI is used in their healthcare as they are to say they are comfortable with how much input they have (63% vs. 21%). Another 16% are not sure.
Widespread Uncertainty About AI's Presence in Medicine
Only a small share of Americans (16%) say their doctors or other healthcare providers have ever used AI in their healthcare. Much larger shares say it has not been used (37%) or that they are not sure (46%)—a striking finding given that AI use in hospitals and doctors' offices is increasingly common.
Adults with higher education levels are more likely to report that their providers have used AI. For example, 25% of those with a postgraduate degree say this, compared with 11% of those with a high school diploma or less education. Yet across education levels, Americans are similarly likely to say they are not sure, ranging from 46% among those with a high school diploma or less to 48% of those with a postgraduate degree.
Among those who say their providers have used AI in their care, the largest share say they understand somewhat well how it has been used (44%). They are more likely to say they understand it not too or not at all well (32%) than extremely or very well (22%).
Chatbots Emerge as a Growing Health Resource
A related Pew Research Center (搜索) report, based on the same survey, examines Americans' use of AI chatbots for health. In all, 34% of Americans ever use AI chatbots for at least one of eight health and medical reasons—a group the Center refers to as chatbot health users.
About a quarter of Americans say they use AI chatbots to diagnose symptoms, and similar shares use them to better understand information from healthcare providers, such as learning more about a doctor's diagnosis or understanding lab results. About three-in-ten U.S. adults (28%) cite getting health information quickly as a motivation, while 22% say they turn to chatbots because they are low-cost or free. Other reasons include helping decide whether to see a doctor (15%) and getting health information they are uncomfortable talking about (18%).
Nearly all chatbot health users find this information at least somewhat helpful: about half (47%) say the information they get is extremely or very helpful, while 48% say it is somewhat helpful. Very few (5%) say it is not too or not at all helpful.
However, these users are split on sharing personal health information with chatbots. They are about as likely to be extremely or very comfortable (29%) as not too or not at all comfortable (26%), with the largest share (42%) saying they would be somewhat comfortable. This comes at a time when some platforms have faced scrutiny for storing and selling people's health data or using it to train large language models, even as some tech companies unveil chatbot tools for healthcare.
Several demographic groups are more likely to be chatbot health users, including Asian Americans, adults under 50, four-year college graduates, and upper-income adults. Overall optimism about AI also plays a role: adults who are more concerned about the increased use of AI are far less likely to be chatbot health users than those who are more excited about these technologies (21% vs. 60%).