Dr GPT: 84% Say ChatGPT Got Their Diagnosis Right

Contributor to Forbes, John Koetsier, discusses the growing trend with American adults relying on AI for medical advice, with 84% supporting the accuracy of AI diagnoses.

According to a recent UserTesting survey involving 2,000 Americans, half of the respondents used large language models (LLM) like ChatGPT for health-related services such as understanding medical diagnoses and getting insights regarding health conditions. Of these, 81% received a diagnosis from the LLM, and a whopping 84% found the diagnosis in alignment with what they received from their physician.

Americans have indicated a higher level of trust towards AI for medical needs as compared to English and Australians. Although some tasks trusted to AI include scheduling appointments or partnering with pharmacies, more than half the respondents indicated they would trust AI for designing treatment plans.

Koetsier shares his personal experience of using AI for interpreting medical terminologies, and deploying ChatGPT for validating a flu diagnosis he self-assessed during the holiday season.

The desire to seek health care solutions beyond the conventional doctor-patient and hospital system has been growing within US citizens. Some 26 million Americans lack health insurance and even for those insured, exorbitant copays or challenges in securing an appointment make alternatives more appealing. Furthermore, lack of local healthcare facilities particularly in rural and inner-city regions creates a pressing need for alternative solutions.

Given the widespread and growing lack of access to medical care, the role of AI will inevitably grow in providing healthcare at the scale needed to match the aging population of the country, according to Lija Hogan, research strategy leader at UserTesting.

The survey also lists several health-related tasks that respondents would entrust to AI, including monitoring sleep patterns, booking doctor's appointments, tracking heart health, diagnosing chronic ailments like heart diseases, coordinating with pharmacies for prescription refills, monitoring blood pressure, diagnosing Covid, providing fertility information, and diagnosing colds and flu.

While many consumers have wearable devices like Apple Watches and Fitbits that can partly fulfil some of these roles, there remains a need for doctors and specific tests for diseases that can only be detected through detailed blood tests or exams beyond the reach of consumer-grade tech.

There are hopes that the best solution might be a united front of doctors and AI, seamlessly delivering health information through wearable tech and smart AI, supported by another defence layer of professionals armed with data and insights from intelligent systems.

The needed connection and seamless integration of healthcare professionals and medically trained AIs is a work in progress.