I'm going to be using Chat GPT a lot more in class. Here's why, and how.

The most popular post on this sub right now is "How to use ChatGPT in any class". I'm here to argue pretty much the opposite, and explain how...and why.<br><br>Along with artifical intelligence, this is going to involve some other tech tools, including speech-to-text. First off, if you're against the use of ChatGPT for any reason, I respect that. If you're looking for ways to integrate this into your grading, read on.<br><br>**The Big Problem**<br><br>The big problem I face, and I think many of us face in this sub, is grading. I teach high school social studies, and each student writes anywhere from 400 to 1200 words a week. I have 300 students. When I first started teaching, I would give a lot of written feedback. Now, I do my best to grade a rubric and give any critical feedback verbally. I would rather take a pay cut than give up my weekends. I know others disagree with this stance, and that's OK.<br><br>**The Big Solution**<br><br>I'm going to be using OpenAI's recently-released Teacher's Tool. With this tool, I'll be able to upload any assignment and run a report on the similarity. This, by itself, isn't new. Quetext and Turnitin offer similar features. However, OpenAI's tool does something different: it's able to make a guess at how much percentage of the writing is written by a human being. This is kind of a big deal.<br><br>**The Bigger Solution**<br><br>You've probably heard of speech-to-text tools. I'm going to be using Otter. As I grade, I'll be having a smart speaker read a student's writing out loud (with a somewhat natural voice), and then the tool will type out what it hears. What this allows me to do is listen to a student's writing --- in their voice --- and then use the rubric as a guide to give verbal feedback.<br><br>For example, if my rubric has a box for "Sentences are often poorly constructed", I'll explain why, with an example from the writing. This, again, isn't revolutionary. Many of us already do this. With this tool, though, I'll be able to listen to what any AI says the student said, and give feedback as I do that.<br><br>So, if I'm grading and I hear something that doesn't sound right, I'll make a note of it. I'll then go back to the writing, and check if what the student actually said is any different from what chatGPT said. If it is, I'm more likely to give the student the benefit of the doubt. I can then use the OpenAI tool to verify my suspicions.<br><br>**Conclusion**<br><br>I'm not going to prioritize "gotcha" policing of AI. I'm not going to focus more on whether or not a student used AI. Rather, I'm going to use AI tools to create a grading system that is:<br>A) more just<br>B) more efficient<br>C) more engaging and interactive<br><br>If I ever find evidence of a student using AI, I'm going to...do nothing. What I care most about is whether or not a student is truly learning, based on what they submit. I'm not going to be cracking down on AI. I'm going to be using new tech tools to improve the quality and fairness of all my grading.