Admittedly, I’m not going to use ChatGPT to write this. One reason is that I need as many writing samples of my own writing that I can get in order to train it to write like me, and the other reason is that I think that would be a little too meta and beyond good taste.
I have been using ChatGPT for three months now, and I want to talk about how it has transformed my work, my education, my teaching, and my life. It really serves several different roles for me, and it’s worth it to document them here.
Work
In my role as a central office administrator, I do a lot of reading and a lot of writing. The writing is typically involves communication, interpretation, and extrapolation. I have found ChatGPT can do a lot of the hard work of my job comparing a situation to policy, comparing an updated policy to an old one, and getting key details from a text or an email.
The most powerful aspect of this for me so far has been the work it did on the qualitative section of a survey that I sent out. Instead of me having to pour over hundreds of qualitative responses and making meaning of them. I had ChatGPT categorize them for me. Of course, I didn’t just blindly copy and paste, and everything I do with ChatGPT always starts off as a conversation. I call him “Greg” in my head. For this situation, in particular, I started off with why I was doing the survey, what the questions were, and what I hoped to learn from it. I didn’t need Greg for the quantitative part, but Greg made the quantitative part happen in a matter of minutes. This saved me hours of work, and it did it more precisely than I would have given the low priority of this work and the time constraints.
Here is what this looks like. First, here is what a typical Microsoft (or Google) Forms spreadsheet looks like:
Microsoft Forms Excel Export
Taking the information from this spreadsheet and turning in into data is easy for the quantitative parts. In fact, Microsoft Forms does that for you anyway. The qualitative parts aren’t quite as easy. In real research, a sophisticated coding process is typically used for this, and I haven’t tried to use ChatGPT for that, but this is a very watered-down version of that. I took all the qualitative responses for an individual question and had Greg look at it.
Here is what happened when I gave this input: “In a survey given to teachers about a new grading scale, they were asked. Have you noticed any positive or negative effects on student motivation since implementing the new grading scale? (open-ended) Please categorize their answers and quantify the results,” and then I pasted the text from the column.
Greg replied, “Categorized answers and their corresponding quantities:
Positive Effects: 10 Negative Effects: 4 Mixed/Neutral Effects: 10 Not Applicable/Not Noticed: 5
The most common type of answer is mixed/neutral effects, with 10 responses falling under this category. Meanwhile, positive effects come in second place with 10 responses, and negative effects come in third with 4 responses. There are also 5 responses indicating that the respondent did not notice any effects or that the question was not applicable to their situation.”
Then I asked, “Can you make me a list of the responses in these categories in a table?” Here is what it gave:
Now, imagine doing this for ten questions like this and then having Greg write a summary. It gave this summary:
“Key Themes in the qualitative responses:
Concern about the long-term effects of the new grading scale: Some respondents expressed concern about the potential long-term consequences of the new grading scale, such as lower expectations for students and a false sense of accomplishment.
Mixed feelings about the benefits of the new grading scale: Some respondents saw the new grading scale as a positive development that gave students more opportunities to succeed, while others were worried that it could lower standards and expectations.
Importance of setting high expectations and holding students accountable: Some respondents emphasized the importance of setting high expectations for students and holding them accountable for their work, even if the grading scale is more lenient.
Differences in opinion about the appropriateness of the new grading scale: There were differing opinions about whether the new grading scale was a good idea or not, with some respondents expressing support and others expressing concern.
Need for clear communication about the grading scale: Some respondents mentioned the importance of clearly communicating the new grading scale to students, parents, and other teachers.
Concern about the impact on college readiness: Some respondents were worried that the new grading scale could give students a false sense of their abilities and preparedness for college.”
This is just one example of many, but this type of analysis work is the real time-saver with Greg. I know a lot of folks use it for email, but I do not like for Greg to write my emails for me, generally, unless they are very low-stakes. I would never send out something heartfelt and personal that was generated by AI. That’s how you wind up in a situation like this. However, some emails that are purely professional with no emotional stakes are perfectly fine for Greg. With high-stakes emails, I like for Greg to look over them for me and to look for bias, unprofessional language, and of course, for any typos. It’s hard to proof your own work, and instead of tying up a colleague to do it, now I have Greg for that (that’s a really poorly-worded sentence, and I wish I hadn’t already boxed myself in by saying I wouldn’t have Greg look at it).
I’ve also noticed when someone “GPTs me” at work. Like, seriously? You had GPT write that to me? It makes me feel some kind of way, but I’m not sure which kind of way yet. I’m mindful of that when I’m using it for email.
Greg has saved me countless hours at work, and I started keeping track of everything I used it for, but it started to be way too much for to keep up with. This is what I had before I quit keeping track of it.
In my next installment in this series, I’ll talk about how I use it to enhance my teaching as an adjunct.