Chatbots such as ChatGPT, Copilot and Gemini have quickly become a part of daily life. In fact, the use of chatbots to find health information is a growing sector of AI. An estimated 16% of Americans say they turn to chatbots with questions about health!

While chatbots may be a useful tool for general health information, it’s problematic to use AI for diagnosing health problems.

Chatbots are designed to sound human by using sympathetic language and providing answers with confidence. That means chatbots may easily gain your trust, even though the information they provide may be inaccurate or misleading.

There have been cases of harmful outcomes for people who have followed health advice from chatbots. This is especially concerning for people who use chatbots for mental health support without also seeing a mental health professional.

Medical and mental health professionals are highly regulated through licensing bodies, and they must continually update their education to keep their license to practice. These professionals are also subject to oversight by licensing organizations. Chatbots do not have such oversight.

In addition, all medical professionals are required to protect their patients’ privacy. Violations of patients’ privacy can lead to fines and other sanctions.

Chatbots are not bound by privacy rules. If you are using generative AI to research health conditions, check the chatbot’s privacy settings first to make sure the chatbot is not storing your conversations or using them to improve their answers for other users. Never share identifying information such as your name and location.

Although chatbots can be useful, it is important to have a primary care provider to help navigate your health concerns. Primary care providers get to know you well, and they can spot signs of problems even before you may be aware of them. Your primary care provider can answer your questions and provide personalized care that meets your unique needs.

It is also important to develop health and media literacy skills to combat misleading information. Verify chatbot answers by checking reliable resources that have a history of providing valid information.

Many areas of the U.S. have limited access to healthcare and mental health providers. Someday, AI may help fill in those gaps, but right now health organizations are warning users that the technology just isn’t ready yet.

Sources:
American Psychological Association
ECRI.org
The Hospital and Healthsystem Association of Pennsylvania