More People Are Asking AI Health Questions, and That’s Good – Inside Sources

New data from Pew Research show a growing number of Americans are using AI chatbots for health reasons. Patients use chatbots to learn about symptoms, diagnoses, treatments and lab results — and even to help decide whether to see a clinician. 

Some will be quick to argue that seeking health or medical advice from a chatbot is dangerous, but an increasing number of people turning to chatbots for health-related information could actually be a good thing for both patients and physicians.

Unfortunately, outdated policies could make it harder for patients to access this information. Even worse, some states are making it harder to use AI in certain health settings, such as mental health care. Lawmakers have considered imposing restrictions on AI, claiming it violates physician licensing requirements. Organized medicine is already urging lawmakers and regulators to require physician oversight of AI that performs clinical functions, using anachronistic licensing laws to preserve physicians’ role as gatekeepers even as AI becomes capable of performing more of those functions.

A better response would be for policymakers to examine how AI, like telemedicine before it, can expand access, and whether states should consider alternatives to the 19th- and 20th-century licensing framework that constrains these technological advances.

It is important to consider why Americans may be turning to AI for their health and medical questions. Cost and help in deciding when to seek treatment were among the top reasons individuals gave for consulting chatbots before medical professionals. Many Americans lack the time or resources to see a doctor for a minor ailment or injury, but they may still want general guidance on what to do and whether to see a clinician. In this way, AI is not replacing medical advice or doctors but helping fill an important gap by providing better, more personalized information.

AI has the potential to reduce the information asymmetry that has long characterized the doctor-patient relationship. This can empower patients to make more informed choices and exercise greater control over their healthcare. That is also why lawmakers should carefully consider the effect that policy proposals restricting the use of AI in medicine or other licensed applications could have on free speech. These restrictions can limit individuals’ access to information they could otherwise obtain from books or other sources while also restricting the expressive rights of AI developers by limiting the information they can use to build their systems.

AI has another potential advantage compared to human physicians: it doesn’t judge. Patients may feel more comfortable sharing sensitive or embarrassing matters with an AI they might hesitate to discuss with another person, and those details can sometimes be crucial to making the right diagnosis. Stigma can be a factor in why individuals, including physicians themselves, do not seek treatment for certain problems like mental health issues.

Of course, AI can make mistakes. So can doctors. Growing evidence suggests that, for some tasks, AI can already perform as well as or better than physicians. In a randomized clinical trial published in JAMA Network Open, GPT-4 alone substantially outperformed physicians on diagnostic reasoning tasks.

We need serious conversations about data privacy and ensuring individuals make informed choices about their medical data. Some individuals may be comfortable with the risk, while others may want to provide more sensitive health information only in anonymized ways or through AI products with more limitations on data use. As we help consumers understand the privacy choices around their medical data when interacting with AI, we should remember that most people who use AI this way say they find it helpful.

As AI improves, it will likely perform more of the tasks requiring physicians and other health professionals, and at a fraction of the cost. That can increase the supply of healthcare services, give consumers more choices, intensify competition with human providers, and put downward pressure on prices.

ChatGPT is not your doctor, but AI chatbots can be an important tool for empowering patients, improving access to information, and even providing better medical care. This future is not guaranteed. Outdated policies, or those that focus only on risks, could prevent the typical American from realizing these benefits.