ChatGPT

 

"What do you think? Ban it, use it..."

"Well, it's certainly going to change how we teach writing."

"It's going to change the landscape of teaching and learning."

"Change..."

"Change..."

"Change..."

Change surprises us sometimes. Though folks have been working on AI and Machine learning for decades, ChatGPT seemed to sneak up on high school teachers out of out of nowhere. Suddenly, nearly sentient some said. It's not sentient, of course; it's not like the robots in Rossum's Universal Robots, by Karel Čapek, nor like the AI companion, Klara, of Klara and the Sun by Ishiguro.

More a tool, ChatGPT buzzed about campuses across the globe with how it worked, what it could do, and how we teachers might use it or detect it if students used it. 

Like many, I played with it. I asked for haikus and sestinas. 

Singapore Sestina #1, ChatGPT

I was tickled with the sestina. The haiku got repetitive quite quickly though.

12 Mushroom Haiku, ChatGPT

Playing around I saw how the tool could support idea generation. I wondered though, how long would it take a student to get dissatisfied with what the bot produced or frustrated by trying to write the perfect command? 

While I appreciated the creative jumpstart, I wondered how it'd write our current task, a This I Believe essay inspired by those shared on National Public Radio. 

With a workshop read around the next day, I started drafting a piece with Chat GPT.

I wish I'd kept a record of the commands I fed the bot, but I didn't and I've cleared the chats so many times by now. 

Needless to say I pieced together a draft. Reread it. Revised it a bit (and kept my revisions in blue). 

I brought the draft to class the next day, clarified the two font colors and sent it around our read-around circle. In that circle we pass drafts of our current pieces.  See all of the drafts here.

With each pass we take 1-2 minutes to read, then silently pass the next draft when time is called. It's quick. We pass for 12-18 minutes and get a decent big picture view, enough of a look-see to talk strengths and next steps, of our writing as a group.  

Once we'd done the read around and reflected together, we talked ChatGPT. How, when or for what would using it fall within our academic integrity guidelines? When would using it violate our academic integrity? 

Students talked thoughtfully about what the bot generated sharing things like "it gives you general ideas," and "it couldn't write the personal parts, only you could." So true!  

Then drafts in hand then left class to get working on their next steps. We regularly use Turn It In . com to scan work for plagiarism in our English dept. The day before revisions were due, I decided to have us play with bot detectors and see what we noticed. I framed our play by wondering aloud about the results my own essay got when I ran it through the detectors. 

We used GPTZero and ChatGPT itself. For ChatGPT we asked it: How likely is it that chatgpt wrote this essay: [copy and paste essay]. The room buzzed as students uploaded their pieces and then read the results and then tried to upload other items and fool the accuracy checkers. We screenshot the checker results and students dropped those screenshots into the GoogleDocs they were working in. 

One student asked ChatGPT to explain perplexity and burstiness scores--that took us down an interesting rabbit hole for a few minutes. Overall I rated that instructional sequence as effective and efficient!  I appreciated students having the time to do it--to check themselves, to inquire into the tools and to talk about what academic integrity means in practice. I enjoyed the facilitator stance, taking on the inquiry --that was decidedly much more pleasant and engaging than investigating after the fact. Win, win.