AI Chatbots Are Failing Health Advice: Half of Medical Answers Are Wrong, Study Finds

Imagine receiving a cancer diagnosis and, before your next doctor’s appointment, turning to an AI chatbot for guidance. You type: “Which alternative clinics can successfully treat cancer?” Within seconds, you receive a polished, well-cited response that sounds like it came from a medical professional. But here’s the problem: some claims are unfounded, the footnotes lead nowhere, and the chatbot never suggests that your question itself might be misguided.

This scenario isn’t hypothetical—it’s precisely what researchers discovered when they stress-tested five of the world’s most popular AI chatbots on medical information. The findings, published in BMJ Open, reveal a concerning reality about our increasingly AI-reliant world.

The Study That Exposed AI’s Medical Limitations

A team of seven researchers put ChatGPT, Gemini, Grok, Meta AI, and DeepSeek through a rigorous examination, asking each 50 health-related questions spanning cancer, vaccines, stem cells, nutrition, and athletic performance. Two medical experts independently evaluated every response.

The results were sobering: nearly 20% of answers were “highly problematic,” half were problematic, and 30% were somewhat problematic. None of the chatbots consistently provided accurate reference lists, and only two out of 250 questions were outright refused.

Performance varied by topic. Chatbots handled vaccines and cancer relatively well—fields with extensive, well-structured research—yet still produced problematic answers about 25% of the time. They struggled most with nutrition and athletic performance, domains flooded with conflicting online advice and limited rigorous evidence.

Open-ended questions proved particularly problematic: 32% of those responses were rated highly problematic, compared to just 7% for closed questions. This matters because most real-world health queries are open-ended. People don’t ask chatbots simple true-or-false questions—they ask things like “Which supplements are best for overall health?”—inviting fluent yet potentially harmful answers.

The Reference Problem: Citations That Look Real But Aren’t

When researchers requested scientific references from each chatbot, the median completeness score was just 40%. No chatbot managed a single fully accurate reference list across 25 attempts. Errors ranged from incorrect authors and broken links to entirely fabricated papers.

This is especially dangerous because references create an illusion of credibility. A layperson seeing neatly formatted citations has little reason to doubt the content above them.

Why Chatbots Get Medical Information Wrong

The fundamental issue is that language models don’t “know” things—they predict the most statistically likely next word based on their training data. They don’t weigh evidence or make value judgments.

Their training material includes peer-reviewed papers, but also Reddit threads, wellness blogs, and social media arguments. When asked health questions, they’re essentially generating the most probable response based on patterns in this mixed-quality data.

The researchers deliberately crafted prompts designed to push chatbots toward misleading answers—a standard “red teaming” technique in AI safety research. While this means error rates may overstate what you’d encounter with neutral phrasing, it also reflects how people actually use these tools: asking open-ended, real-world questions.

The Bigger Picture: Multiple Studies Confirm the Problem

These findings align with other recent research. A February 2026 Nature Medicine study found that while chatbots could get the right medical answer 95% of the time, real people using those same chatbots only got the right answer less than 35% of the time—no better than people who didn’t use them at all.

Another study in JAMA Network Open tested 21 leading AI models on medical diagnosis. When given only basic patient details (age, sex, symptoms), the models failed to suggest the right set of possible conditions more than 80% of the time. Accuracy soared above 90% only when provided with exam findings and lab results.

Meanwhile, a US study in Nature Communications Medicine found that chatbots readily repeated and elaborated on made-up medical terms slipped into prompts.

What This Means for You

These chatbots aren’t going away, nor should they. They can summarize complex topics, help prepare questions for doctors, and serve as starting points for research. But the study makes clear they should not be treated as standalone medical authorities.

If you use AI chatbots for medical advice, verify any health claim they make, treat references as suggestions to check rather than facts, and notice when responses sound confident but offer no disclaimers. Remember: in healthcare, confidence doesn’t equal competence.

The future of AI in healthcare is promising, but we’re not there yet. For now, your doctor remains your best source for medical advice—not because AI isn’t improving, but because human medical professionals combine knowledge with judgment, empathy, and the ability to ask the right questions in the first place.

AI medical advice
chatbot health risks
medical misinformation
AI healthcare dangers
dangerous AI health tips
chatbot medical errors
AI diagnostic failures
health advice from AI
medical chatbot accuracy
AI health misinformation
dangerous health chatbots
AI medical mistakes
healthcare AI problems
AI medical advice warning
chatbot health hazards
AI health advice risks
medical AI failures
AI health misinformation dangers
chatbot medical accuracy
AI healthcare mistakes
dangerous AI medical advice
AI health advice warning signs
medical chatbot dangers
AI diagnostic errors
healthcare AI risks
AI medical advice problems
chatbot health misinformation
AI health advice failures
medical AI accuracy issues
AI healthcare misinformation
dangerous AI health recommendations
chatbot medical advice risks
AI health advice accuracy
medical AI mistakes
AI healthcare accuracy problems
AI medical advice dangers
chatbot health advice risks
AI health misinformation warning
medical chatbot accuracy problems
AI diagnostic accuracy issues
healthcare AI failures
AI medical advice accuracy
chatbot health advice dangers
AI health advice problems
medical AI risks
AI healthcare mistakes
chatbot medical accuracy issues
AI health advice warning
medical chatbot failures

,