January 23, 2026 – Synthetic intelligence (AI) chatbots in healthcare high the checklist of the highest well being know-how threats for 2026, based on an annual report from ECRI. Chatbots that depend on massive language fashions (LLMs) – resembling ChatGPT, Claude, Copilot, Gemini and Grok – produce human-like and expert-sounding responses to consumer questions. The gadgets should not regulated as medical gadgets nor have they been validated for healthcare functions, however are more and more utilized by docs, sufferers and healthcare professionals, based on ECRI. But greater than 40 million folks flip to ChatGPT for well being info every single day, based on a current evaluation by OpenAI.

ECRI says chatbots can present helpful help, however can even present false or deceptive info that may result in vital affected person hurt. ECRI subsequently advises warning when utilizing a chatbot for info that would impression affected person care.

As a substitute of actually understanding context or that means, AI methods generate responses by predicting strings of phrases based mostly on patterns realized from their coaching knowledge. They’re programmed to sound assured and all the time present a solution to fulfill the consumer, even when the reply just isn’t dependable. ECRI consultants say chatbots have advised incorrect diagnoses, advisable pointless checks, promoted poor medical provides and even invented physique components in response to medical questions, all whereas sounding like a trusted professional.

Learn extra from ECRI.