63% Followed Unverified AI Health Advice, 58% of Millennial Parents Did So for Children, 42% Misrepresented Information for Privacy Reasons

By Talkerresearch | Reviewed by Irfan Ahmad

One in seven Americans trust AI chatbots as much as they trust a doctor, according to new research.

A survey of 1,000 millennial parents and 1,000 general population Americans explored how they are using AI to manage their own health or their children’s health, and whether they understand how AI generates the advice it provides.

Image: Betterimagesofai - CC BY

Confidence in AI is growing

Among Americans who use AI for health purposes, the number of respondents who trust AI as much as they trust a doctor nearly quintuples compared to non-AI users, and 60% said they would believe AI-generated health advice even if it contradicted guidance from a healthcare provider.

The survey conducted by Talker Research for MD Live by Evernorth found that nearly two-thirds (63%) of AI users admit they have followed health advice from AI without first confirming it with a healthcare provider.

That confidence may stem from the way AI delivers information: Half of respondents said that if AI gives them a clear or reassuring response about their health, they’re likely to accept it without questioning it.

Those who use AI for health purposes ask an average of nine health-related questions each week.

These respondents say they rely on it because it’s conveniently available at any time (37%), faster than seeing a doctor (31%) and provides reassurance about their health concerns (30%).

However, 28% of respondents say they only verify the answer they receive from AI less than half the time.

The trend extends to decisions about children

Among millennial parents who use AI for health purposes, 58% admit they have acted on AI-generated health advice for their children without first confirming it with a healthcare provider.

Many also said they would be comfortable solely relying on AI for a range of decisions involving their children’s health, including evaluating minor symptoms (80%), determining what type of care is needed (71%) and seeking medication guidance (71%).

Despite the confidence in AI, Americans often overlook how their input may affect the accuracy of the guidance it provides. Nearly eight in ten respondents (78%) did not recognize that the way a question is worded influences the answer AI provides.

More than half (53%) admitted they have reworded or adjusted their health question when they didn’t like the first answer they received, including 27% who say they always do so.

Many AI users also ask questions in ways that steer AI toward a desired response.

Thirty-eight percent said they describe symptoms and ask what could be wrong, while about a third ask whether a symptom could be connected to a condition they already suspect they have (34%) or whether a specific diagnosis could explain what they’re experiencing (32%).

“The survey indicates that people are often going to AI already hoping for a certain answer, whether that’s reassurance or confirmation of a fear,” said Dr. Maggie Williams, medical director of virtual primary care at MD Live by Evernorth. “That’s very different from how a doctor works through a symptom with a patient, asking questions, gathering context and ruling possibilities in or out before reaching a conclusion.”

Many overestimate what AI actually knows

The survey also found that many Americans overestimate how AI works in the first place.

One in five respondents believe that AI only pulls information from trusted medical sources (22%), and the same share of regular users believe AI factors in their personal medical history even when that information has never been provided.

At the same time, many respondents acknowledge they do not provide AI with a complete picture of their health when seeking guidance.

Forty-two percent admit they’ve even lied about who they’re submitting a health question about for privacy reasons.

More than three-quarters of regular users (76%) leave out their medical history, demographic details (76%) and personal information due to privacy concerns (81%).

As a result, many AI users report receiving answers that weren’t fully relevant to their circumstances.

Nearly six in 10 (59%) said they later realized health information they received from AI didn’t fully apply to their situation, and nearly half (49%) say they acted on that information anyway before discovering it wasn’t the right fit.

Human oversight remains important

Despite growing adoption of AI for health purposes, respondents expressed concerns about receiving incorrect advice (34%) and a lack of human oversight (31%).

One in four said they plan to use AI for more healthcare decisions in 2027, while 27% said they would trust AI-generated information more if it is reviewed by a doctor.

“The results show people will continue to turn to AI for their health care questions — and having people more engaged in their health is a good thing,” said Dr. Williams. “But healthcare providers add important clinical guidance by catching details a patient may not consider relevant, asking follow-up questions and putting the information into the context of that person’s health history. As we consider a patient’s full care journey, it’s important to understand how AI and physicians can come together to provide the clearest picture to patients while ensuring appropriate safeguards and clinical oversight.”

Read next: Study Finds 80% Missed Chatbot Errors Despite Participants’ Basic Knowledge of the Subject

by External Contributor via Digital Information World

Study Finds 80% Missed Chatbot Errors Despite Participants’ Basic Knowledge of the Subject

By Radboud University | Reviewed by Irfan Ahmad

A chatbot that speaks naturally is more likely to be perceived by users as intelligent and useful, even if it spouts nonsense, research by communication scientists at Radboud University shows. An eloquent AI system that shares incorrect information is rated just as highly by users as a curt AI system that provides correct answers.

Image: Solen Feyissa - Unsplash

Most people are by now quite familiar with ChatGPT, Gemini, Claude and other chatbots: every day, people ask millions of questions and receive neatly formulated answers in return. Yet these language models still tend to run wild with so-called ‘hallucinations’: unreliable, dubious answers that are presented with great certainty as the truth. Communication scientist Marianna de Sá Siqueira and her colleagues investigated the effects of elegant language use on the credibility of chatbots.

Nonsense in fluent language

For the experiment, participants were shown one of four variants of a conversation with a chatbot about The Lord of the Rings. Two variants featured fluent sentences and human-like language, one of which gave correct answers and the other did not. In addition, there were two variants with short, businesslike answers, one of which provided entirely correct information whilst the other presented (subtly) fabricated facts.

Unsurprisingly, users were most positive about the chatbot with elegant language and correct answers. However, a chatbot that writes fluently but spouts factual nonsense was rated by participants as just as intelligent and pleasant as a chatbot that communicates in a stilted manner but does provide factually correct information. It also emerged that as many as 80 per cent of participants did not notice the errors at all, even though they had been selected in advance for their basic knowledge of The Lord of the Rings.

"Above all, it shows how easily people can be led astray and how quickly we come to trust something that is actually spouting nonsense," says Barbara Müller, associate professor and co-author of the study. "If a text sounds elegant and natural, people automatically attribute intelligence to it. You then subconsciously tend to think: this is fine, I don’t need to check it any further."

Blind trust

The research highlights the danger of blind trust in artificial intelligence. Professor and co-author Tibor Bosse says: "So-called language models such as GPT and Claude are primarily trained to produce statistically plausible and easy-to-read sentences. These systems do not always take the truth very seriously. They are designed to sound human and convincing, but whether the information is factually correct is of secondary importance."

"People used to complain that chatbots were annoying and knew nothing. Now they sound so clever that we’re too quick to believe them," says Bosse. According to him, this cannot be solved with a simple disclaimer at the outset. "Most chatbots show a single line of text at the start stating that the output may not be entirely accurate, but those companies are certainly not going to indicate of their own accord with every answer that the system is uncertain. If we really want to protect users from chatbots spouting nonsense, we need to, for example, oblige tech companies to make that uncertainty visible. If a chatbot itself expresses doubt, you prevent people from taking nonsense as the truth."

Literature reference:

de Sá Siqueira, M. A., Bosse, T., & Müller, B. C. N. (2026). Users’ perception of chatbot language fluency and hallucination. Computers in Human Behavior: Artificial Humans, 10, 100412. https://doi.org/10.1016/j.chbah.2026.100412.

Read next: Deceptive Google Ads Generated Over 100 Million Impressions Despite Existing Safeguards, Some Running for Years

by External Contributor via Digital Information World