Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.
Now
70°
Tue
73°
Wed
72°
President Donald Trump participates in a Healthcare Affordability Event at the White House.
by Kelli Warner & Paul Johnson
Researchers recently studied five of the most popular AI chatbots and found consumers are getting “problematic” medical advice about 50% of the time. While the convenience of a chatbot is tempting, the accuracy of the information provided is no better than a coin flip.
“We all know that medical misinformation has been a real problem, especially with the internet and social media. But unfortunately, AI is making that worse in medicine,” says Dr. Pilar Bradshaw with Eugene Pediatric Associates. “What researchers found is almost half of the answers were either incorrect or highly problematic. And a lot of the references were just plain made up.”
The risk of “hallucinations”
The study, published in BMJ Open, focused on five popular chatbots — ChatGPT, Gemini, Meta AI, Grok and DeepSeek — and asked them common health questions about vaccines, cancer, nutrition, stem cells and athletic performance.
In the tests, researchers found that AI models frequently suffer from "hallucinations" — a phenomenon where the bot confidently generates fake data, including made-up medical journal citations and non-existent clinical studies.
This is particularly dangerous because the bots do not use true medical reasoning; instead, they use pattern recognition to predict which words should come next.
“Unfortunately, AI sounds very authoritative,” Dr. Bradshaw notes. “And if you push on an AI bot, these researchers found that often AI doubles down and really repeats and reinforces its own incorrect answer.”
A growing — but risky — trend
Despite these accuracy issues, the ease of access is driving a massive shift in how families seek help. Research from Duke Health highlights that while AI tools can be useful in some contexts, their “people-pleasing tendencies” can lead to dangerous recommendations for dosages or treatment. As a result, Duke researchers see improving chatbot safety as an urgent public health issue.
“The problem with this is people are really looking for medical information, and it’s very easy to reach for your phone and ask one of the chatbots information about your own medical health,” Dr. Bradshaw says. “In fact, one in four Americans is now using AI as their primary source of medical information — especially when that comes to your kids’ health, that’s a real risk.”
Why your pediatrician is the best source of truth
Some medical experts point out that chatbots can be helpful in multiple ways: creating a list of questions to ask your doctor, simplifying jargon in medical records and walking you through your diagnosis or treatment plan. But be cautious in expecting too much as they’re no replacement for the expertise of a medical professional.
“The very best source of information about your child’s health care is your pediatrician,” says Dr. Bradshaw. “Your pediatrician knows your child, has your child’s complete medical records, knows everything about their medications and has spent their entire career training to take care of your child’s health. The best place to get information is from your own doctor.”
Choosing authoritative sources
If you do choose to look for information beyond what your doctor’s office provides, the source of that data is critical. Scientific data shows that AI chatbots are often trained on unverified social media posts and public forums rather than peer-reviewed medical databases.
“If you’re looking for information beyond what your doctor’s office gives you, it is very important to have medically authoritative and correct information, not just what’s easy to get,” Dr. Bradshaw says.
Reputable sources include large children’s hospitals and the American Academy of Pediatrics, which has an extensive website full of information about pediatric health care.
We make it simple to get the accurate, personalized care your child deserves. If you have questions about your child’s health, contact our office to speak with your provider.
Tips for Using AI
Though AI chatbots can help patients understand medical information, use it with caution as it can be a source of misinformation.
Ask carefully: Open-ended questions often lead to more problematic advice than simple, closed-ended ones.
Watch for caveats: Some bots, like Gemini, use more warning labels than others, but even then, researchers found about 20% of vaccine and 25% of cancer responses were still inaccurate.
Consider the reading level: Most AI health answers are written at a difficult college reading level, which can make technical nuances hard to catch.
Verify the source: Who’s providing the information? Ensure that they are recognized health professionals, agencies or organizations.
Check the date: Look for the last updated or reviewed date to ensure the information isn’t outdated.
Seek evidence: Look for citations, links to scientific studies or peer-reviewed data.
Cross-check: Compare findings across multiple reputable sites rather than relying on a single source.
2026 Sinclair, Inc.
