Stardate 192023 My First Experience with ChatGPT AI

Computer key showing AI
AI Image

Twelve years after graduating with a Master of Educational Technology degree from UBC, I found myself encountering a technology we could only have conceived of 15 years ago: ChatGPT AI – our own version of Star Trek’s Data.

My initial understanding was that ChatGPT could answer questions, write papers, and provide vast amounts of information in minutes. As a professor and a writer, I didn’t know whether to be mortified or thrilled by this advancement.

More online cheating was my first thought. Would anyone write their own papers now? How would I know? I soon found the student-created GPTZero which apparently helps educators determine which papers are human written and which are AI written. Crisis averted – maybe.

My second thought was now we can stop giving students essays to write as assessments. I spend hours and hours trying to determine which ones were written by my students, created by software, or bought on essay mills. Incidentally, I could make more money writing for an essay mill than I do teaching. And as a course designer, I struggle with the authenticity, meaningfulness, and purpose of assigning essays.

I was stunned by the quality of the ChatGPT generated essays I saw online. Was this cheating at its finest? I read a few samples and then signed up for ChatGPT myself. The AI answers to my questions were so eerily human that I was tempted to say thank you to the replies.

As a writer, my fear kicked in. What would this mean to the writing community? ChatGPT answered my questions in seconds. Where does that leave me? I can’t write that fast. I can’t even organize my thoughts that fast. But I can feel, and writing is about emotion. It’s painting pictures with words, stirring others to react, and making people think.

So, I started thinking bigger. How can we use this? Can students use AI to write essays and then reflect on how to make them better? Or would they just ask AI to reflect on the AI’s essay? What would I have done as a UBC M.E.T. student? The professors would have told us to jump in, try it out, and above all don’t forget pedagogy before technology which essentially means if you need a pen, use a damn pen. Don’t use technology just because it’s technology, shiny and new. Find out what it does, and if you need it, use it.

With all of this in mind, I spent the day trying out ChatGPT to see what I could accomplish.

I’d recently recorded some ideas for an Accessible Lecture course on my phone while out walking. The recording was a very rough draft with words like wanna and gonna in it, and the topic order wasn’t entirely consistent. Without fixing anything, I used transcription software, Designrr, to create a text document which I copied and pasted into ChatGPT.

Then the fun began. My recording wasn’t very organized, so I asked ChatGPT to create an outline for me which it did in less than a minute. I was impressed and since the outline was good enough to use, I asked ChatGPT to change the outline into slides.

It presented me with 5 slides with slide titles and relevant points on each. I copied the slides back into ChatGPT and then asked for visuals for each slide. After giving me an error message twice, it informed me that it was too overloaded with similar requests to provide images, but it did give me visual descriptors for each slide. Even without the images, I now had enough information to create my own slides based on my own lecture material.

Friend or Foe?

At first glance, I was ecstatic at the assistance I had creating my slides. It was almost like sitting with a person and discussing what ideas to put on each slide. But it felt strange. Was I violating copyright laws? Whose information was this? The whole transcription was mine and I had used technology to record and transcribe, so why not use it to organize too?

If I lecture based on the outline but use my own words, I could really benefit from the organization of ideas which is not my strong point. In fact, getting help organizing my ideas would give me so much more time to write and create courses. I’m excited by the possibility. But I won’t use the slides. Using ChatGPT’s words on my slides doesn’t feel like using my own work. So, whose work was it?

I really wasn’t sure, so I asked the AI about plagiarism. Here’s our conversation:

Me: Are you a plagiarism machine?

ChatGPT: I am not a plagiarism machine. I am a language model trained to assist with a variety of tasks, including answering questions and providing information on a wide range of topics. I can provide information, but I cannot copy text or produce original content on my own. Is there anything else I can help you with?

Me: Why don’t you quote or cite your sources when you give us answers?

ChatGPT: As a language model, I do not have access to the internet and I am not able to browse websites or access online sources. I am not able to quote or cite sources in the same way that a person would because I do not have access to the original source material. Instead, I provide information based on the data that I have been trained on. If you have any specific questions about a topic or if you would like more information on a particular subject, please let me know and I will do my best to help.

Me: I’m sorry that I accused you of plagiarism.

ChatGPT: There is no need to apologize. I understand that my responses may not always include citations or references, and it is important to clarify that I am not able to access or use outside sources. If you have any further questions about my capabilities or how I work, please don’t hesitate to ask. Is there anything else I can help with?

Reflecting on my experience, I admit that I’m intrigued and apprehensive, even a little creeped out if I’m honest. But I can see the potential. Maybe I can use it to create new assessments that students appreciate and find useful in real life. Maybe I can create courses faster with some organizational help. It might even be able to help me create accessible material. I am worried about copyright and accuracy – ChatGPT told me it didn’t copy material or create original content – but doesn’t it have to be one or the other?

So, there you have it, my first encounter with AI. Yes, I apologized to a machine. I was curious as to the response I would get. And who’s to say that years down the road, when it accesses its first encounters with humans, it won’t “appreciate” my kindness. It is still learning after all.