From AP News: With hundreds of millions of people turning to chatbots for advice, it was only a matter of time before tech companies began offering programs specifically designed to answer health questions.
In January, OpenAI introduced ChatGPT Health, a new version of its chatbot that the company says can analyze users’ medical records, wellness apps and wearable device data to answer health and medical questions. Currently, there’s a waiting list for the program. Anthropic, a rival AI company, offers similar features for some users of its Claude chatbot.
Both companies say their programs, known as large language models, aren’t a substitute for professional care and shouldn’t be used to diagnose medical conditions. Instead, they say the chatbots can summarize and explain complicated test results, help prepare for a doctor’s visit or analyze important health trends buried in medical records and app metrics.
The article suggests five points to consider before using a chatbot to analyze your health
1. Chatbots can offer more personalized information than a Google search
Doctors and researchers see AI platforms like ChatGPT Health as an improvement over the status quo, providing more personalized and specific information than a Google search. While not perfect, these chatbots can offer useful information when used responsibly and with as much context as possible.
2. If you’re having worrisome symptoms, skip AI
Seek immediate medical attention for symptoms like shortness of breath, chest pain, or severe headache. Approach AI programs with skepticism, especially for major medical decisions.
3. Consider your privacy before uploading any health data
AI bots offer benefits from users sharing personal medical information, but this data isn’t protected by HIPAA, the federal privacy law governing sensitive medical information. With this in mind, OpenAI and Anthropic claim to keep health information separate and subject to additional privacy protections, requiring user opt-in and allowing disconnection.
4. Testing shows chatbots can stumble
AI chatbots struggle with human interaction despite their ability to accurately identify medical conditions. Communication issues, such as incomplete information from users and unclear responses from chatbots, can affect their effectiveness.
A 1,300-participant study by Oxford University Foundation that AI chatbots don’t improve health research decisions over people doing their own research and using personal judgment. However, the study was conducted in 2024 and it did not use the latest chatbot versions, including new offerings like ChatGPT Health.
5. A second AI opinion can be helpful
Chatbots could improve by asking follow-up questions and eliciting key details from users. Consulting multiple chatbots can increase your confidence in the information provided.
ChatGPT Health can help you navigate everyday questions and spot patterns over time, so you feel more informed, prepared, and confident for important medical conversations. pic.twitter.com/UK6U4OosDn
— OpenAI (@OpenAI) January 7, 2026
Read more at AP News
Never miss an episode. Subscribe to the show by downloading The DML News App or go to Apple Podcasts.
The Dennis Michael Lynch Podcast archive is available below, with the most recent on top.
