60% of Americans Trust AI Medical Advice Over Their Doctors, Survey Finds

A new survey reveals that 60% of Americans who use AI for health-related questions would trust its answers even when they contradict their doctor's advice. Meanwhile, 63% have already acted on AI-generated medical guidance without consulting a healthcare professional.

The findings come from a survey conducted by Talker Research and published in recent days, highlighting the growing reliance on artificial intelligence for medical information and the potential risks of accepting its responses without verification. The survey was conducted last summer among 2,000 adults, including 1,000 millennial parents and 1,000 respondents from the general population. In each group, half of the participants used artificial intelligence for health-related questions.

Parents Increasingly Rely on AI for Children's Health

The findings are particularly concerning when it comes to children's health. Among parents who use AI chatbots for medical questions, 58% reported following the technology's advice about their children without first consulting a doctor.

On average, respondents who use AI for health information ask nine medical questions per week. The most common reasons for turning to the technology were convenience (37%), receiving answers faster than scheduling a doctor's appointment (31%), and seeking reassurance (30%). Half of these users said they would accept an answer that seemed clear or reassuring without questioning it, while another 28% acknowledged checking the accuracy of AI-generated responses in fewer than half of their interactions.

The survey also revealed that 59% of respondents had subsequently discovered that information provided by AI did not fully apply to their particular medical circumstances. Nearly half (49%) had already acted on the advice before realizing the discrepancy.

How Questions Are Worded Can Influence AI Responses

Another significant finding concerns users' understanding of how artificial intelligence works. According to the survey, 78% of respondents did not realize that the wording of a question could influence the answer they received. Furthermore, 53% admitted rephrasing a question after receiving an initial response they did not like, while 42% acknowledged providing inaccurate details about the person they were asking about in order to protect their privacy.

These behaviors raise questions about the reliability of AI-generated medical guidance, particularly when users provide incomplete or misleading information or seek answers that reinforce their existing expectations.

The respondents' primary concerns about AI health advice were inaccurate information (34%) and the absence of human oversight (31%). More than a quarter (27%) said they would trust AI-generated answers more if a physician reviewed them. Interestingly, 22% believed that artificial intelligence obtains its information exclusively from reliable medical sources.

Why Doctors Still Play an Essential Role

"The survey shows that people often turn to AI when they are already hoping for a particular answer," said Dr. Maggie Williams, medical director at MDLive, the company that commissioned the research.

Williams emphasized that the findings indicate people will continue using artificial intelligence to seek health information, making professional medical guidance increasingly important. According to Williams, healthcare professionals provide essential clinical judgment because they can recognize details and symptoms that patients may not realize are relevant to their condition.

While AI can provide convenient access to general health information, the survey highlights the risks of relying on automated responses without professional evaluation, particularly when making treatment decisions or addressing children's medical needs.

It is important to note that the findings were based on participants' self-reported behavior rather than independently verified medical outcomes. The survey was also commissioned by MDLive, a company that provides physician-supervised online healthcare services and therefore has a commercial interest in the subject.

These limitations should be considered when interpreting the results. Nevertheless, the findings raise important questions about how people use AI for medical advice and whether they understand the potential consequences of trusting its answers without consulting a qualified healthcare professional.