Archive · 1 Jul 2025
Study: AI Health Advice Can Be Inaccurate and Dangerous
AI chatbots, like the ones you might use to search for information, can sometimes give the wrong answers when it comes to your health.
A new study showed that these chatbots often give out incorrect medical advice that could be dangerous.
This could lead you to think you are sick when you are not, or to get the wrong treatment.
Experts say we should be careful about using AI for health advice and that these AI tools should be checked very carefully before being used to help people.
Study warns against using AI chatbots for health advice.
Researchers found chatbots can provide inaccurate and misleading medical information.
Four out of five AI chatbots generated disinformation in most responses.
AI use for health advice can lead to missed diagnoses and incorrect treatments.
Experts recommend thorough testing of AI systems before widespread healthcare use.
- Who
- People who use or consider using AI chatbots for health information.
- What
- Study findings on the accuracy of AI chatbots providing health advice.
- When
- Information not provided in the articles.
- Where
- Findings are globally applicable.
- Why
- To warn against the potential dangers of using AI for medical advice.
- Key Finding
- AI chatbots often give inaccurate medical advice.
- Study Publication
- Annals of Internal Medicine
- AI Systems Analyzed
- OpenAI, Google, Anthropic, Meta, X Corp
- Percentage of False Responses
- Up to 100% for some AI systems
- Risk
- Misdiagnosis, delayed or incorrect treatment
