By Nate MacLeitch

Nearly half of UK adults admit to using online health information to self-diagnose, often because they can’t get a timely NHS appointment. Conversational AI facilitates a similar impulse: About one in six US adults already use mainstream chatbots like ChatGPT for health advice at least monthly.

When untrained for medical purposes, these “Dr. GPT” habits fuel anxiety, misdiagnoses, and compliance headaches for providers.
Anxiety is a growing concern and the most common mental health disorder in the US. In 2025, roughly 42.5 million US adults are living with an anxiety disorder. Turning to tools like Dr. Google and Dr. GPT to seek reassurance and answers can have an adverse effect if conversations are not monitored carefully. These tools are trained on the entire internet, from medical blogs to Reddit posts written by the general public. When the bots respond with certainty and confidence, based on misinformation, users can find themselves misdiagnosed.

This risk underscores the need for careful deployment, because while the same tools can mislead, they also hold remarkable promise. AI medical assistants are showing vast potential in widening access to care and reducing the gap between complex medical language and public understanding. Healthcare providers who leverage these tools safely will improve their reliability and patient outcomes.

The deliverables of Dr. GPT
We are hearing revolutionary stories now of chatbots accurately diagnosing cancer patients after their doctors misdiagnosed them. While this use case is still uncommon, it shows the level of care patients are already beginning to receive with these tools by their side.

A patient might tell Dr. GPT: “I have the worst headache of my life.” And Dr. GPT returns a list of potential suggestions of what the patient could be experiencing:

● Migraine
● Tension headache
● Cluster headache

The patient self-diagnoses and takes some ibuprofen for the pain. In theory, this saves multiple unnecessary visits to general practices and emergency rooms, alleviating patients and doctors.

AI is proven to deliver general medical advice, education, and support for understanding conditions, treatment options, and preventive care strategies. It is great at explaining medical terms in plain language, breaking down communication barriers between doctors and patients, and helping them better follow their treatment plans.

Patients can ask things like, “Why is it important for my condition to take this medication before/after my meals?” or, “If I don’t feel sick, why should I keep taking my pills?” A 2021 meta-analysis of individuals with type 2 diabetes found that those with good medication adherence (≥80%) experienced 10% fewer hospitalizations and 28% lower all‑cause mortality. Giving them the tools to better understand their treatment plans can help encourage them to stick to them.

Health IT Homepage