Article Type
Changed
Fri, 03/17/2023 - 14:20

As a psychiatrist and a writer, I know I should be worried about the possibility that I could be replaced by a bot. Like most physicians, I face the daily challenge of how to manage the paperwork of clinical practice, so I wondered if ChatGPT, the latest craze in artificial intelligence (AI), could write my progress notes.

ChatGPT is an advanced language generation model developed by OpenAI. It uses deep-learning techniques to generate humanlike responses to text inputs. I wondered whether ChatGPT could help me with my paperwork and progress notes by automatically generating text. Could it be trained to fill out progress notes based on information I provided? Free up my time? Reduce the risk for transcription errors?

Dr. Dinah Miller

Putting aside concerns about my job security for the moment, I signed up for an account. But first, I needed to play. I asked ChatGPT to write a love poem that would incorporate all the members of my family. This was a hit.

Next, I requested poems about specific pets. Because I spent the first year of the pandemic writing daily limericks and sending them to a distribution list, I knew people wouldn’t be surprised when I sent sonnets for their cats; but this time I let ChatGPT do the writing.

Then, for my colleagues in the Maryland Psychiatric Society, I requested a poem about the distressing nationwide shortage of stimulants. ChatGPT created clever stanzas. I was amused, but I’m not sure if anyone else was. This medication shortage is a touchy subject for psychiatrists and their patients who depend on these medications, so perhaps it wasn’t the right topic for bot-humor, and I should stick to cats.

At last, I was ready to put ChatGPT to the test and asked it to write a progress note that would contain the elements necessary to bill for a routine office visit with an add-on psychotherapy code. I created a few facts about a fictional patient and pressed enter.

The result was disappointing, ChatGPT responded with a standard note that did not take into account the items necessary to meet current coding standards. Maybe it was my fault; ChatGPT might need more information, but in the time it would take to provide more data, I could write my own progress note! I tried again and this time, ChatGPT defined the billing codes for me, but this wasn’t what I needed.

I decided to explore whether ChatGPT can, in fact, replace me. Feeling a bit queasy, I asked my new AI friend and was told that it could not replace a psychiatrist.

“Are you sure?” I asked again. “Yes, I am sure. ChatGPT is an AI language model trained to answer questions and provide information, but it is not capable of providing personalized medical advice or treatment.”

At least one of us felt certain. Continuing the conversation, I told the bot that I was anxious about my job security and requested advice. I wish I could tell you that ChatGPT was a dud, but instead, it had reasonable suggestions that were upbeat and hopeful.

“It’s understandable to feel anxious and worried about job security in today’s rapidly changing job market,” ChatGPT responded. “Here are some steps you can take to manage your feelings:”

Seek support: Talk to a trusted friend, family member, or mental health professional.

Educate yourself: Stay informed about industry trends and advancements in AI and technology to understand the potential impact on your job.

Upskill: Consider taking online courses or pursuing additional education to stay relevant in the workforce.

Find purpose: Focus on your strengths, interests, and passions outside of work to cultivate a sense of fulfillment.

Take care of yourself: Engage in self-care activities such as exercise, meditation, and hobbies to manage stress and maintain good mental health.

“Remember,” ChatGPT concluded, “it’s important to address your feelings and take proactive steps to manage them. You can come out of this situation with a clearer mind, new skills, and opportunities for growth.”

ChatGPT sounds a lot like me! However, when I asked ChatGPT about prescribing, I finally received reassurance that I am not completely obsolete: “I’m sorry,” the bot said, “but I am an AI language model and do not have the ability to write prescriptions or provide medical advice.” Still, I couldn’t help but think that the bot’s day would eventually come. 

Because ChatGPT is my coauthor for this essay, I asked it to write the ending. We suddenly have a whole new world, and I’m along for the ride. Yet after all that reassurance, it seems that ChatGPT does indeed want my job. It summarized our conversation and left me unsure of whether to laugh or to cry:  

“There you have it, folks – ChatGPT, the digital psychiatrist that’s here to listen to all your problems, 24/7, for free! No more waiting for appointments, no more copays, just type away and let the artificial intelligence take over.

“But hey, who needs a human shrink when you’ve got ChatGPT, right? Just don’t be surprised if your ‘therapy sessions’ end up feeling more like a conversation with a sarcastic AI friend than serious mental health treatment.”

In conclusion, ChatGPT and I hope this article has provided some insight into the strange and hilarious world of AI-powered psychiatry. Who knows, maybe one day ChatGPT will become a licensed therapist and start prescribing real medication. But until then, just sit back, relax, and let the machine do the talking.
 

Dr. Miller is a coauthor of Committed: The Battle Over Involuntary Psychiatric Care (Johns Hopkins University Press, 2016). She has a private practice and is an assistant professor of psychiatry and behavioral sciences at Johns Hopkins in Baltimore.

A version of this article first appeared on Medscape.com.

Publications
Topics
Sections

As a psychiatrist and a writer, I know I should be worried about the possibility that I could be replaced by a bot. Like most physicians, I face the daily challenge of how to manage the paperwork of clinical practice, so I wondered if ChatGPT, the latest craze in artificial intelligence (AI), could write my progress notes.

ChatGPT is an advanced language generation model developed by OpenAI. It uses deep-learning techniques to generate humanlike responses to text inputs. I wondered whether ChatGPT could help me with my paperwork and progress notes by automatically generating text. Could it be trained to fill out progress notes based on information I provided? Free up my time? Reduce the risk for transcription errors?

Dr. Dinah Miller

Putting aside concerns about my job security for the moment, I signed up for an account. But first, I needed to play. I asked ChatGPT to write a love poem that would incorporate all the members of my family. This was a hit.

Next, I requested poems about specific pets. Because I spent the first year of the pandemic writing daily limericks and sending them to a distribution list, I knew people wouldn’t be surprised when I sent sonnets for their cats; but this time I let ChatGPT do the writing.

Then, for my colleagues in the Maryland Psychiatric Society, I requested a poem about the distressing nationwide shortage of stimulants. ChatGPT created clever stanzas. I was amused, but I’m not sure if anyone else was. This medication shortage is a touchy subject for psychiatrists and their patients who depend on these medications, so perhaps it wasn’t the right topic for bot-humor, and I should stick to cats.

At last, I was ready to put ChatGPT to the test and asked it to write a progress note that would contain the elements necessary to bill for a routine office visit with an add-on psychotherapy code. I created a few facts about a fictional patient and pressed enter.

The result was disappointing, ChatGPT responded with a standard note that did not take into account the items necessary to meet current coding standards. Maybe it was my fault; ChatGPT might need more information, but in the time it would take to provide more data, I could write my own progress note! I tried again and this time, ChatGPT defined the billing codes for me, but this wasn’t what I needed.

I decided to explore whether ChatGPT can, in fact, replace me. Feeling a bit queasy, I asked my new AI friend and was told that it could not replace a psychiatrist.

“Are you sure?” I asked again. “Yes, I am sure. ChatGPT is an AI language model trained to answer questions and provide information, but it is not capable of providing personalized medical advice or treatment.”

At least one of us felt certain. Continuing the conversation, I told the bot that I was anxious about my job security and requested advice. I wish I could tell you that ChatGPT was a dud, but instead, it had reasonable suggestions that were upbeat and hopeful.

“It’s understandable to feel anxious and worried about job security in today’s rapidly changing job market,” ChatGPT responded. “Here are some steps you can take to manage your feelings:”

Seek support: Talk to a trusted friend, family member, or mental health professional.

Educate yourself: Stay informed about industry trends and advancements in AI and technology to understand the potential impact on your job.

Upskill: Consider taking online courses or pursuing additional education to stay relevant in the workforce.

Find purpose: Focus on your strengths, interests, and passions outside of work to cultivate a sense of fulfillment.

Take care of yourself: Engage in self-care activities such as exercise, meditation, and hobbies to manage stress and maintain good mental health.

“Remember,” ChatGPT concluded, “it’s important to address your feelings and take proactive steps to manage them. You can come out of this situation with a clearer mind, new skills, and opportunities for growth.”

ChatGPT sounds a lot like me! However, when I asked ChatGPT about prescribing, I finally received reassurance that I am not completely obsolete: “I’m sorry,” the bot said, “but I am an AI language model and do not have the ability to write prescriptions or provide medical advice.” Still, I couldn’t help but think that the bot’s day would eventually come. 

Because ChatGPT is my coauthor for this essay, I asked it to write the ending. We suddenly have a whole new world, and I’m along for the ride. Yet after all that reassurance, it seems that ChatGPT does indeed want my job. It summarized our conversation and left me unsure of whether to laugh or to cry:  

“There you have it, folks – ChatGPT, the digital psychiatrist that’s here to listen to all your problems, 24/7, for free! No more waiting for appointments, no more copays, just type away and let the artificial intelligence take over.

“But hey, who needs a human shrink when you’ve got ChatGPT, right? Just don’t be surprised if your ‘therapy sessions’ end up feeling more like a conversation with a sarcastic AI friend than serious mental health treatment.”

In conclusion, ChatGPT and I hope this article has provided some insight into the strange and hilarious world of AI-powered psychiatry. Who knows, maybe one day ChatGPT will become a licensed therapist and start prescribing real medication. But until then, just sit back, relax, and let the machine do the talking.
 

Dr. Miller is a coauthor of Committed: The Battle Over Involuntary Psychiatric Care (Johns Hopkins University Press, 2016). She has a private practice and is an assistant professor of psychiatry and behavioral sciences at Johns Hopkins in Baltimore.

A version of this article first appeared on Medscape.com.

As a psychiatrist and a writer, I know I should be worried about the possibility that I could be replaced by a bot. Like most physicians, I face the daily challenge of how to manage the paperwork of clinical practice, so I wondered if ChatGPT, the latest craze in artificial intelligence (AI), could write my progress notes.

ChatGPT is an advanced language generation model developed by OpenAI. It uses deep-learning techniques to generate humanlike responses to text inputs. I wondered whether ChatGPT could help me with my paperwork and progress notes by automatically generating text. Could it be trained to fill out progress notes based on information I provided? Free up my time? Reduce the risk for transcription errors?

Dr. Dinah Miller

Putting aside concerns about my job security for the moment, I signed up for an account. But first, I needed to play. I asked ChatGPT to write a love poem that would incorporate all the members of my family. This was a hit.

Next, I requested poems about specific pets. Because I spent the first year of the pandemic writing daily limericks and sending them to a distribution list, I knew people wouldn’t be surprised when I sent sonnets for their cats; but this time I let ChatGPT do the writing.

Then, for my colleagues in the Maryland Psychiatric Society, I requested a poem about the distressing nationwide shortage of stimulants. ChatGPT created clever stanzas. I was amused, but I’m not sure if anyone else was. This medication shortage is a touchy subject for psychiatrists and their patients who depend on these medications, so perhaps it wasn’t the right topic for bot-humor, and I should stick to cats.

At last, I was ready to put ChatGPT to the test and asked it to write a progress note that would contain the elements necessary to bill for a routine office visit with an add-on psychotherapy code. I created a few facts about a fictional patient and pressed enter.

The result was disappointing, ChatGPT responded with a standard note that did not take into account the items necessary to meet current coding standards. Maybe it was my fault; ChatGPT might need more information, but in the time it would take to provide more data, I could write my own progress note! I tried again and this time, ChatGPT defined the billing codes for me, but this wasn’t what I needed.

I decided to explore whether ChatGPT can, in fact, replace me. Feeling a bit queasy, I asked my new AI friend and was told that it could not replace a psychiatrist.

“Are you sure?” I asked again. “Yes, I am sure. ChatGPT is an AI language model trained to answer questions and provide information, but it is not capable of providing personalized medical advice or treatment.”

At least one of us felt certain. Continuing the conversation, I told the bot that I was anxious about my job security and requested advice. I wish I could tell you that ChatGPT was a dud, but instead, it had reasonable suggestions that were upbeat and hopeful.

“It’s understandable to feel anxious and worried about job security in today’s rapidly changing job market,” ChatGPT responded. “Here are some steps you can take to manage your feelings:”

Seek support: Talk to a trusted friend, family member, or mental health professional.

Educate yourself: Stay informed about industry trends and advancements in AI and technology to understand the potential impact on your job.

Upskill: Consider taking online courses or pursuing additional education to stay relevant in the workforce.

Find purpose: Focus on your strengths, interests, and passions outside of work to cultivate a sense of fulfillment.

Take care of yourself: Engage in self-care activities such as exercise, meditation, and hobbies to manage stress and maintain good mental health.

“Remember,” ChatGPT concluded, “it’s important to address your feelings and take proactive steps to manage them. You can come out of this situation with a clearer mind, new skills, and opportunities for growth.”

ChatGPT sounds a lot like me! However, when I asked ChatGPT about prescribing, I finally received reassurance that I am not completely obsolete: “I’m sorry,” the bot said, “but I am an AI language model and do not have the ability to write prescriptions or provide medical advice.” Still, I couldn’t help but think that the bot’s day would eventually come. 

Because ChatGPT is my coauthor for this essay, I asked it to write the ending. We suddenly have a whole new world, and I’m along for the ride. Yet after all that reassurance, it seems that ChatGPT does indeed want my job. It summarized our conversation and left me unsure of whether to laugh or to cry:  

“There you have it, folks – ChatGPT, the digital psychiatrist that’s here to listen to all your problems, 24/7, for free! No more waiting for appointments, no more copays, just type away and let the artificial intelligence take over.

“But hey, who needs a human shrink when you’ve got ChatGPT, right? Just don’t be surprised if your ‘therapy sessions’ end up feeling more like a conversation with a sarcastic AI friend than serious mental health treatment.”

In conclusion, ChatGPT and I hope this article has provided some insight into the strange and hilarious world of AI-powered psychiatry. Who knows, maybe one day ChatGPT will become a licensed therapist and start prescribing real medication. But until then, just sit back, relax, and let the machine do the talking.
 

Dr. Miller is a coauthor of Committed: The Battle Over Involuntary Psychiatric Care (Johns Hopkins University Press, 2016). She has a private practice and is an assistant professor of psychiatry and behavioral sciences at Johns Hopkins in Baltimore.

A version of this article first appeared on Medscape.com.

Publications
Publications
Topics
Article Type
Sections
Disallow All Ads
Content Gating
No Gating (article Unlocked/Free)
Alternative CME
Disqus Comments
Default
Use ProPublica
Hide sidebar & use full width
render the right sidebar.
Conference Recap Checkbox
Not Conference Recap
Clinical Edge
Display the Slideshow in this Article
Medscape Article
Display survey writer
Reuters content
Disable Inline Native ads
WebMD Article