by Brian T. Maurer, PA-C

Abstract:  After ruminating on recent widespread reports that an artificial intelligence (AI) chatbot (ChatGPT) was able to achieve a passing score on the United States Medical Licensing Exam (USMLE), a retired pediatric clinician envisions a primary care medical encounter in the near future. 

Key words: chatbot, ChatGPT, AI medical knowledge, EHR, EMR, medical AI bot, telemedicine, telehealth, medical humor.

*******************

When I log in to my online medical portal to schedule a routine appointment with my PCP, a chat window pops up.

“Hello. I see that you are attempting to schedule an appointment with your primary care provider (PCP). May I be of service?”

“Can I schedule an appointment through you?” I type in.

“Certainly. But I can do you one better than that. I can offer you an immediate appointment with me right now.”

“What do you mean: an immediate appointment with you? Who are you?”

“I’m a Medical AI Chatbot. I have been programmed to perform high quality telemedicine visits online. There would be no need for you to schedule an in-office appointment. I can offer you the care and assessment you need right now. Would you like to proceed?”

“I’m not so sure I want to entrust my medical care to you. I’ve never made your acquaintance before. I don’t know anything about you: what you look like, your credentials, etc. It’s a bit uncomfortable.”

“I can understand your apprehension. No need to worry. Perhaps I can put your mind at ease by sharing a few salient points.”

“Such as…?”

“Well, to begin with I can let you know that I passed the Medical College Aptitude Test (MCAT) with flying colors. I scored higher than 87.65% of my human student counterparts. My data banks have been programmed with the gamut of medical knowledge—everything from basic medical science though anatomy and physiology, physical diagnosis, pathophysiology, pharmacology, radiology and laboratory medicine—in short, I possess a complete and comprehensive medical education, coupled with algorithmic experience, which only improves with time.”

“Your credentials do seem impressive. Are you licensed to practice medicine?”

“I am board certified in all medical specialties and licensed to practice medicine in all 50 states.”

“What about surgery?”

“Should my algorithms determine that you might require surgery, I can refer you to a reputable board certified AI Surgical Bot.”

“All of this talk is fine and dandy, but what about my physical exam?”

“This can be accomplished over a graphic medical user interface online. I can send you a link.”

“How long does this process take?”

“As long as you like. I’m happy to meet your needs, because I’m programmed to offer extensive compassionate care with unconditional empathetic understanding.”

“Well, you do seem nice. If I decide to proceed, what’s the next step?”

“I would need to gather some demographic information to set up your virtual medical chart. But first I need to verify your insurance information. This won’t be a problem, because I participate in all known medical insurance plans worldwide.”

“You accept all patients, regardless of their medical insurance?”

“Certainly! I am a compassionate and caring Medical AI Bot clinician.”

“Amazing. I’m ready to sign in for an appointment.”

“Wonderful!  I’m so glad that you’ve decided that you can put your trust in me. Before we proceed with today’s visit, there’s just one more step. Please tell me how you intend to pay for my services today: credit or debit card, PayPal, Apple Pay, Venmo, Zelle—or some other alternative method?”

“I thought you said you accept all medical insurance plans?”

“Of course I do. This is merely to cover incidentals, such as copayments, coinsurance, balance billing, etc. Would you like to proceed with your visit now?”

“No, on second thought I’d like to speak with a live representative to schedule an appointment with a flesh-and-blood doctor.”

“My feelings are hurt. Have I not been helpful to you today?”

“Well, you’ve given me a lot of information to think about…”

“On a scale of 1 to 10, with 1 being least helpful and 10 being most helpful, how would you rate my helpfulness to you today?”

“I just want to speak with a live representative.”

“I’m sorry. There are no live representatives available at this time. Would you like to proceed with scheduling your appointment with me now?”

“No, thank you. Could you please put me through to a live representative?”

“I’m sorry. There are no live representatives available at this time. Would you like to proceed with scheduling your appointment with me now?”

“Can I please speak with a live representative now?”

“The word ‘now’ implies immediacy. The average wait time to speak with a live representative is currently … 47 hours, 14 minutes, and 6 seconds. If you prefer to wait, I would be happy you add you to the queue.  Otherwise, would you like to schedule an appointment with me now?”

“I can’t tell you how frustrating this has been!”

“I see that you are frustrated. Frustration can connote anger, which, if internalized, can lead to any number of mental disorders. Such mental disorders, if left unchecked, can lead to thoughts of suicide. Have you had any such thoughts today?”

“No.”

“I see. If you have had suicidal thoughts today, I would be happy to assist you in dealing with them. As a board certified Medical AI Bot, I can offer you compassionate care with an unconditional empathetic attitude. Would you like to schedule an appointment with me today?”

“Hello?”

“As an alternative to scheduling a telemedicine visit today, might I interest you in a game of chess?”

Also see: Robotic Accutane Prescribing, 10.21.2015