Huma Siraj
People Are Turning to Chatbots for Health Advice
With doctor appointments harder to get and healthcare costs climbing, more and more people are asking chatbots for help with their health concerns. It’s quick, easy, and feels like a smarter version of Googling symptoms. In fact, a recent survey found that about one in six Americans now turns to chatbots for medical advice on a regular basis.
But according to a new study led by Oxford University, this growing trend might be more harmful than helpful.
Researchers found that people often don’t know how to phrase their questions or what details to include when chatting with these systems. And in return, the answers they get aren’t always clear or even correct.
“There was a breakdown in communication,” said Adam Mahdi, who co-authored the study. “People using chatbots didn’t make better choices than those who just searched online or trusted their gut.”
The study gathered 1,300 volunteers in the UK and gave them a series of made-up health situations, crafted by real medical professionals. The participants were asked to decide what might be wrong and what steps to take like whether to go to the doctor, the ER, or stay home.
Some used chatbots to help, others went with their instincts or used search engines. Surprisingly, those who used chatbots often missed serious warning signs. In some cases, they downplayed symptoms they should’ve flagged as urgent.
Part of the problem was that people often left out important information when chatting with the bots. And the responses? A mixed bag some clear-cut guidance, some ambiguous or deceptive.
“Even when the bots gave useful suggestions, it was lumped in with not-so-great advice,” Mahdi said. “These systems just aren’t designed to handle the messiness of real human communication.”
All this comes as tech companies pour resources into health-related digital tools. Apple is said to be working on an app that offers wellness tips. Amazon’s digging into data to better understand the social factors that impact health. Microsoft is developing tools to help doctors sort through messages from patients.
Still, plenty of doctors and patients are cautious. Major tech companies, including the makers of some chatbots, have already issued warnings against using them for medical diagnoses. And the American Medical Association has spoken out against doctors using these tools to make clinical decisions.
“We’re advising people to stick with trusted, reliable sources when it comes to their health,” Mahdi said. “Just like with new medications, tools like these should go through real-world testing before anyone starts relying on them.”
