Picture this. You've got a weird rash, a two week wait for your GP, and a portal message system that responds with the energy of a Victorian postal service. What do you do? Increasingly, you ask a chatbot.
So is that actually happening? According to a new commentary in Annals of Internal Medicine from Robert Shpiner, a critical care doctor at UCLA, yes, and at scale. Recent data show 32% of US adults used AI for health information or advice in the past year.
Why? Not because they think Dr Chatbot is better than you. Because it's instantly available, and everything around it isn't. Shpiner points to the actual reasons patients give: no timely appointments, no regular physician, cost, and a general lack of faith the health system will respond at all. Billing headaches, prior authorisation faff, and portals that feel designed by someone who hates you all push people further toward the AI tab.
His point isn't 'ban the chatbots' or even 'warn people harder about the risks.' It's that governance and safety warnings are rearranging deckchairs if the actual reason patients are leaving the ship goes unaddressed. Fix the appointment bottlenecks and communication mess, and the chatbot traffic is a symptom, not the disease.
Bottom line: next time a patient turns up having already consulted an AI, don't read it as distrust in you specifically. Read it as a two week wait they weren't willing to sit through in silence.
Can't get an appointment, can't get through on the portal, can't get a straight answer on your bill. No wonder the rash asked the robot first.