One in three Americans use artificial intelligence chatbots for medical advice—that’s according to a study by KFF, a health policy organization.
France Bélanger, professor at Virginia Tech’s Pamplin College of Business, said this could have positive benefits. “You’re going to get more information. Often, you can have it translated in terms you can understand. And then that can really help you talk to your doctor.”
But relying too much on AI for medical advice is problematic. Bélanger warned these chat bots are trained to speak as though they are experts.
“The systems do not know you. And they haven’t seen you,” Bélanger explained. “So when you say something, they say ‘I understand your problem.’ Well that’s not true. The system doesn’t understand your problem. Or ‘you’re right.’ Or ‘I feel sorry to hear this.’ When the system says ‘I feel sorry to hear this,’ that’s not true. The system doesn’t feel sorry for you. It’s the way they’re trained to answer.”
Bélanger is also concerned about privacy issues because information shared with an AI bot could be leaked or sold to outside companies.
“And what people often do not realize— it’s not HIPPA protected,” Bélanger said.
Overall, Bélanger said AI has a lot of potential to help the medical industry. But when it comes to medical advice, only your doctor and nurses know your context and can effectively diagnose your situation. She cautioned that AI should not replace those relationships.
