#Dispatches Returns: How are you navigating ChatGPT in the classroom? 

After a hiatus, #Dispatches From the Frontlines returns! This series crowdsources questions of interest to our followers. From the wild to the weird, we ask you–and you deliver!

This week’s question – How are you navigating ChatGPT in your teaching and assignments?

We all see that AI sidebar story on every website and battle its encroachment into our apps, but what we haven’t seen is how you all are managing it in your classrooms. What’s a professor to do—give in, or give out?

Read on for our readers’ takes. (Some comments have been adjusted for length.)

Find next week’s question and form to share your response here! (we will continue with the current Q for one more week.)

We’re taking this on by sorting responses into three (flexible) categories:

1. Saying NO to ChatGPT. Limit or lose it.
2. Saying MAYBE to ChatGPT–on certain conditions
3. Saying YES to ChatGPT– This is where we are. Accept it, or even embrace it!

• Saying NO to ChatGPT. Limit or lose it.

I rewrote all of my in-class syllabi so that almost all writing is in class. Bad experiences with AI. (Tenured full professor, Humanities)

 

I’ve lost patience with AI generated answers, so my policy (as described in the syllabus) is that if I believe that an assignment was generated by AI the assignment gets a zero. Then it is up to the student to contact me and explain how they came up with their answer if they want the grade changed. Usually this results in a student explaining that they “wrote the answer but just put it in grammarly to make it sound better” and then I ask them to re-do the assignment without grammarly. (Assistant Professor, Social Sciences)

 

I’ve altered assignments so they are not easily completed using AI, or at least, they can’t completed well enough to earn more than a B. (Non-tenure track professional)

 

I ban student use of it because I need to be able to assess a student’s skills/competencies. Even with a ban I’ve been bogged down last semester with suspicious writing and AI detection. I’ve tried to use detectors, but they aren’t reliable. I feel I have to triangulate to try to have evidence so I can impose penalties as needed. It isn’t fair to the students who actually do the work if folks who simply use AI are getting away with it. I include a survey quiz that highlights the course policy, but still folks were using it. I’m finding a mix of students: some clearly using it to cut corners, save time, because they are overwhelmed, and a more ambiguous group using it for editing and translating. I fear for my ESL students who may be using translating software that is being flagged as AI/is AI and who also don’t have the self-advocacy or outside supports to navigate the system if they do get in trouble. For myself, I’ve been using it to help with updating some in-class activities. I think it’s most useful for experts in a field who can critically assess it’s quality and accuracy. For students who are learning in a field, they don’t know how to do that. (Tenured, public health/health sciences)

 

I don’t use it for ethical reasons and lack of trust. When students use it their papers are crap. They hand up a polished turd that has a lot of long words but misses the point and uses irrelevant or out of date references. I would never get AI to write my student feedback since I want the feedback to be useful and specific so they get better at what they are doing. In terms of AI writing, why should I bother reading what no one has bothered writing? (Masters student, teaching education)

 

I followed my university’s Academic Integrity once for an egregious use of AI. The entire process consumed a significant amount of time and involved me trying to explain how it was entirely obvious to me that the student’s work was AI-generated. (Our university’s AI-detection software registered the paper as 100% AI-generated, but I still had to explain myself.) I simply don’t have time to go through this process. Instead, I assign short, specific, and challenging writing assignments and grade them very attentively. AI-generated work typically scores a 1 or 2 out of 10 points. I find it most effective to get out of the AI-detection and accusation game and just give all written work the point value it deserves. (Assistant Professor, Humanities)

 

• Saying MAYBE to ChatGPT–on certain conditions.

I am using the institution’s policy and have information on each rubric. I talk with my students about how to use AI to help study and to help you look up words or terms you don’t know, but I do not allow it on assignments. The rubric guides what I’ll do. (Professional staff / adjunct, Arts/Music/Theater)

 

I’m teaching my students to use it to help them brainstorm and imagine new curricular design. They first generate an AI unit plan (or they can also use a plan they find online, in Pinterest, etc) and critique the AI/existing draft. Then conduct/add their own research to complicate the ideas and make them more sophisticated. Finally, they expand their ideas into their own research-informed full curricular unit — where they CAN use AI to template their lessons (but most of them find it’s not good at this task). I’ve found that AI is *terrible* at writing unit and lesson plans that aren’t straight lecture, but there’s often a few good ideas in among the bad ones that can be built up/out. Because my students are brand-new teachers, critiquing the bad ideas and surfacing/building on the good ones are both productive, and it helps them feel like they aren’t starting completely blind in a brand-new professional genre. (Tenured prof, teaching education)

 

I’m allowing chatgpt and google translate with 2 conditions: first that they cite it in a bibliography (for intellectual integrity) and second that they include 1-2 sentences about what they learn from using it. Examples I gave were that if they use AI tools to correct their French, they should note what kinds of mistakes they are prone to making. I’m hoping this will make AI more of a tool than a crutch, but we’ll see ? (Tenured prof, humanities)

 

I have embraced that AI will be with us forever, and have encouraged students to use AI for assignments, with the caveat that they fill out a one-page (5-minute) template. The template asks which tool they used, how they used it (e.g. brainstorming, summarizing something, etc), how they fact-checked it, and how they cited it. This makes it more of a contract with students, where they understand plagiarism directly from AI won’t be tolerated, without ignoring that AI is going to be used a lot in everything from here on out. (Tenured, social sciences)

 

• Saying YES to ChatGPT. This is where we are. Accept it–even embrace it! 

I slowly but steadily started embracing or rather accepting the use of GenAI tools both for my own work (e.g. adapting texts for specific language proficiency level for my language courses, brainstorming ideas for in-class activities, but mostly – seeking editing assistance when writing in English) and showing students the benefits but also the “perks” of these tools in the class. i have adopted a “moderate” AI use policy in my classes: I allow it for brainstorming, language assistance, outlining, but I also require students to be transparent about its use. (NTT, Humanities)

 

I have never and will never use [ChatGPT] to do my actual job, but I have zero issues with people who use it to cut down on endless busywork–since we have been doing that using templates for years anyway. Since I don’t teach my courses in English, my students learn from me that ChatGPT is fairly advanced in English, but highly unreliable in other languages so far. It’s a good opportunity to talk about cultural influence, privilege and power. It’s also a reminder that using it for our class would result in a hot mess given the algorithms are still toddlers for us. (Assistant Professor, Humanities)

 

I teach technical communication for an undergrad major, graduate major, minor, and certificate program. It’s very important that my students understand and know how to use AI. I believe it will be crucial in their future careers to be able to efficiently and effectively use AI technologies. I have redesigned each of my assignments for both my undergrad classes and my graduate classes to specifically implement strategies for them to use AI in those assignments. Then I have been explicit about assignments that they can’t use AI, For example I told them they can’t use AI to create their discussion board posts on online class boards. I explicitly say in those instructions do not use AI to write this. I want to be sure that students are reading the articles and engaging with the material. In those instructions I say it is important for graduate students to know how to read multiple articles, synthesize them, summarize them, and put them in conversation with one another. That’s why AI is not allowed for that assignment. However, for their professional and More reader Focused assignments, I encourage them to use AI noting that the assignment should not sound like AI when it’s done. No one wants to read something that sounds like a robot wrote it. But it’s critical that my students know how to use this technology and create a text that sounds like a human wrote it. I am embracing AI as much as I can because I recognize that it is a helpful tool for me and my students. (Tenured, humanities)