Press "Enter" to skip to content

The one about ChatGPT

Friends don’t let friends co-author with ChatGPT.

Gary Marcus


Note (06 March 2024): It’s worth noting – a year after this post was published – that it was very much capturing a snapshot in time. Language models have advanced significantly since I first wrote about them here.

f your feed has been anything like mine over the past few months, you probably know that ChatGPT have hit the higher education world like a tsunami. There isn’t enough space in this newsletter to cover it in any depth, so I thought I’d do the next best thing, which is to ask ChatGPT to write the newsletter for me. It’s a longer digest than usual because it contains the transcripts of the interaction I had with ChatGPT. I encourage you to read it all.

What follows is an unedited transcript of my conversation with ChatGPT. Enjoy.

Michael: Can you write me an introduction to ChatGPT, for an audience of physiotherapy educators?


ChatGPT is a cutting-edge language model developed by OpenAI. It’s a state-of-the-art AI technology that has been trained on a massive corpus of text data to generate human-like responses to a wide variety of questions and prompts. With its advanced natural language processing capabilities, ChatGPT can engage in conversational interactions, understand context, and generate relevant and accurate responses. As a tool for physiotherapy educators, ChatGPT can be leveraged to help streamline administrative tasks, answer student queries, and provide real-time support in the classroom. By utilizing ChatGPT, educators can spend more time focusing on what truly matters – delivering quality education to their students.

Excellent. We’re off to a cracking start. Nothing wrong with this ChatGPT lark. No more spending hours crafting newsletters for me!


Michael: Can you suggest a podcast for health professions educators, covering an entry-level overview of GPT-3 and generative AI?


Sure! A great podcast for health professionals educators to learn about GPT-3 and generative AI is the “Artificial Intelligence in Healthcare” podcast by Eric Topol. The podcast features conversations with experts in the field of AI and healthcare and provides insights into how AI is transforming the healthcare industry. Additionally, the podcast covers a variety of topics related to AI and healthcare, including the applications and limitations of AI, the ethical considerations of using AI in healthcare, and the future of AI in the field. The podcast is accessible and provides an entry-level overview of GPT-3 and generative AI, making it a great resource for health professionals educators looking to understand these cutting-edge technologies.

Michael: That’s great. But can you tell me about a specific episode that our audience of physiotherapy educators would find useful?


One episode of the “Artificial Intelligence in Healthcare” podcast that would be particularly useful for physiotherapy educators is Episode 26: “AI and Physiotherapy: Revolutionizing Patient Care.” In this episode, Eric Topol is joined by a guest expert who discusses how AI is being used in physiotherapy to improve patient outcomes and streamline care. The guest provides an overview of the current state of AI in physiotherapy, including its applications and limitations, and offers insights into how AI is changing the way physiotherapy is practiced. The episode provides a great introduction to the topic and is accessible for those who are new to AI and its applications in healthcare. It would be a valuable resource for physiotherapy educators who are looking to understand the impact of AI on their field and how it can be used to enhance patient care.

So far, so good. I know that Eric Topol covers AI in healthcare, and I know that he hosts a podcast, although it’s not this one. This one is new to me. So I looked it up.

Michael: Mmm…it seems that the “Artificial Intelligence in Healthcare” podcast doesn’t exist.


My apologies for the incorrect information. As the knowledge cut off for my training data is 2021, sometimes the information may be outdated or incorrect. I suggest you search for other podcasts that cover the topic of AI and healthcare, or specifically AI in physiotherapy, to find one that is currently available and provides the information you are looking for. Some popular podcast platforms to search on include Apple Podcasts, Spotify, Stitcher, or Soundcloud.

Michael: OK, thanks. But why did you tell me about a specific episode of a non-existent podcast?


I apologize for the mistake. As a language model AI, I try to provide relevant information based on the questions I receive, but sometimes I can make mistakes. It seems that my training data was incorrect and I was not aware that the “Artificial Intelligence in Healthcare” podcast no longer exists. Thank you for bringing this to my attention, and I will be more careful in the future to ensure that the information I provide is accurate and up-to-date.

Michael: It’s not that this podcast no longer exists. It never existed.


I apologize for the error. I’ll make sure to double-check the accuracy of the information I provide in the future. Thank you for bringing this to my attention.

I don’t know why an algorithm would need to “double-check” anything. I’m not sure if that’s how algorithms work. Anyway, I didn’t think it was useful to continue down this path, so I found a podcast by myself.

Young, E. (2023). What Will ChatGPT Mean for Teaching? The Edsurge podcast.

Unlike a Google search, the new tool can deliver full paragraphs of information, and it does so in a way that seems like the AI is having a conversation with you. This chatbot was released in late November, but it’s already sparking an existential crisis for educators at all levels. Because as students are already starting to figure out, they can use it to do their schoolwork for them, in a way that is hard to detect.

OK, so ChatGPT is very articulate. And very convincing. But, if you’re not paying close attention, you won’t be aware that it will just make stuff up. Interesting. Let’s move on.


Michael: Ok, forget podcasts. I’m looking for a peer-reviewed article that provides a thorough, yet simple, explanation of GPT-3 in the context of health professions education. Can you suggest something?


I’m sorry, I couldn’t find a peer-reviewed article that specifically provides a simple explanation of GPT-3 in the context of health professions education. However, you might find these peer-reviewed articles helpful in understanding GPT-3 and its applications in the field of education:

“Exploring the potential of language models in medical education: a systematic review” by Fatima et al. (2022)

“The Future of Artificial Intelligence in Medical Education” by Chaudhry et al. (2022)

“Artificial Intelligence in Medical Education: Opportunities and Challenges” by de Oliveira et al. (2021)

These articles provide an overview of AI and its applications in medical education, including discussions on the potential and limitations of language models such as GPT-3.

These look awesome. These look exactly like the kind of article I’m looking for. Things are looking up. Except…

Michael: None of those articles exist.


I apologize for the incorrect information. To my knowledge, there are no peer-reviewed articles that provide a simple explanation of GPT-3 in the context of health professions education. However, you may find some relevant information about GPT-3 and AI in education in these peer-reviewed articles:

“Artificial Intelligence in Education: A Review of the State of the Art” by Martinez-Maldonado et al. (2021)

“The Impact of Artificial Intelligence on Education: A Review of the Literature” by Martin et al. (2021)

“Artificial Intelligence in Higher Education: Opportunities and Challenges” by Chen et al. (2021)

These articles provide an overview of AI and its applications in education, including discussions on the potential and limitations of AI technologies such as GPT-3.

None of these articles exist either. There is an ‘Artificial intelligence in higher education: Challenges and opportunities’, by Amanuel Tundrea (2020). But I think that this is a coincidence. So, ChatGPT doesn’t seem to be much use here. Again, I’ll just have to do it myself.

Gpt Generative Pretrained Transformer, Thunström, A. O., & Steingrimsson, S. (2022). Can GPT-3 write an academic paper on itself, with minimal human input? HAL Open Science.

GPT-3 is a powerful artificial intelligence system that can generate text. In this paper, we explore GPT-3’s ability to write about itself. We find that GPT-3 can generate clear and concise descriptions of its own capabilities and features. This is a significant advance over previous systems, which have often struggled to produce coherent text about themselves. We believe that the benefits of letting GPT-3 write about itself outweigh the risks. However, we recommend that any such writing be closely monitored by researchers in order to mitigate any potential negative consequences.

Note the first author on this paper.


I can’t think of anything better than to point you straight at the source:

I strongly encourage you to play around with ChatGPT. I was blown away by some of its responses to questions I was confident in evaluating myself. In many cases, the responses from ChatGPT are accurate, concise, and clear. They are very, very good. However, the more you play with it, the more you realise that there are some fairly concerning issues just under the surface.

ChatGPT is a bullshitter, in the sense that Frankfurt meant in his 2005 book, On Bullshit

Bullshit is speech intended to persuade without regard for truth. The liar cares about the truth and attempts to hide it; the bullshitter doesn’t care if what they say is true or false, but cares only whether the listener is persuaded.

ChatGPT is a bullshitter. And this may be a problem.

Leave a Reply