How Teachers Are Using ChatGPT in Class (Opinion)

Save to favorites
Print

Email Facebook LinkedIn Twitter

Copy URL

Larry Ferlazzo

Opinion Contributor, Education Week

Larry Ferlazzo is an English and social studies teacher at Luther Burbank High School in Sacramento, Calif.

Today’s post is the latest in an ongoing series about artificial intelligence in the classroom.

‘In Its Infancy’

Mary Beth Hertz teaches high school students art and technology in the Philadelphia school district, has been at the forefront of integrating technology into the classroom for over a decade, is a published author, and can be found at http://marybethhertz.me:

My friend and colleague Bonnee Bentum and I love to dig deep into tech tools and explore their possibilities in our classrooms. Back in late January, Bonnee popped by my room, and we began to riff about the release of ChatGPT. When we heard about the public release of ChatGPT-3, we knew we had to get our students involved as soon as possible.

We know that our students here in Philadelphia often face disadvantages when it comes to technology access. We were dismayed to hear about N.Y.C. public schools blocking ChatGPT. Our fly-by conversations in the hallway often turned into longer dialogues around the implications of the fact that, for some students, generative AI tools would be more likely to be viewed as “cheating,” while in more affluent districts and schools, students would be leveraging it as a leg up into career and industry applications. Our dialogues began to morph into sharing ways that we were using AI with our students.

Despite the controversy about ChatGPT spreading through the education world, we explored the use of AI tools with our students. When ChatGPT came out, we had already been discussing artificial intelligence and its inherent biases, including the work of former Google AI ethicist Timnit Gebru.

Students learned how artificial intelligence and large-language models “learn” from the data they are trained on. I shared a biography that ChatGPT generated about me, highlighting parts that were completely fabricated to show some of the limitations of the tool. I had students converse with ChatGPT and reflect on the conversation. One student who asked it for dinner recipes noted that it didn’t ask about allergies and stated that “they pull information from things that they already know about, which kind of gives us humans the upper hand to correct certain issues that are being shown.”

I also had students in my entrepreneurship class use it to simplify mission statements and language for their business pitches. Our school network’s IP configuration made ChatGPT think we were a bot when many students tried to connect at once, which blocked our access, so we defaulted to using poe.com or you.com, which integrate ChatGPT.

Bonnee’s senior English class collaborated with Paul Allison from YouthVoices to annotate sections of the 1619 Project docuseries. Through YouthVoices’ built-in AI Mojo tool, programmed to build essays from annotations and edit them for grammar, sentence structure, and spelling, students saw how their annotations could lead to more long-form writing. However, they also learned the limitations of the technology and felt more urgency to review and revise their writing to maintain their own voice in their writing.

While using another tool, NowComment, students used the built-in AI “thought partners” to have a specific kind of conversation with the AI. For instance, students could ask the AI to “be snarky” or “give helpful feedback and advice.” As Bonnee reflected, “How often do we see students writing because they want to write?” Students were writing not just to fulfill an assignment but because they found the process engaging and exciting. One of her students, who is interested in exploring a career in tech, told her, “I never thought I would have a teacher who would teach me AI.”

The use of generative AI in K-12 settings is complex and still in its infancy. We need to consider how these tools can enhance student creativity, improve writing skills, and be transparent with students about how generative AI works so they can better understand its limitations. As with any new tech, our students will be exposed to it, and it is our task as educators to help them navigate this new territory as well-informed, curious explorers.

Transparency

Paul Wilkinson became a secondary English and social studies teacher in 2011 after careers in journalism, law, and government. Contact him via https://paulwilkinson.com:

Our district blocks AI for students, so I’ve used it on their behalf to create readily achievable challenges to help them learn content and practice skills and give them feedback to understand that content and master those skills. I personally pay a $20 monthly subscription for GPT-4, and use it with Canvas, which our district released to teachers recently to prepare for its broader adoption this fall.

A preface: I am transparent with students to model ethical standards for using AI. They watch me generate AI content in class, know they get annotated feedback in Canvas using AI, and watch over my shoulder as I submit their texts to GPT4 for analysis and comments on their work while feedback slowly crawls onto the screen. My students understand there’s no way one teacher could give this much feedback to 165 students without large-language-model help, and the novelty of getting such detailed feedback improves engagement.

First, I used it to write texts and questions based on our content and standards. For example, I input our vocabulary and standards for our financial literacy unit and asked it for “curriculum to allow 7th graders to master this content.” A conversation followed in which AI and I created narratives with 7th grade characters, multiple-choice questions, and detailed explanations of particular topics. I reviewed everything for accuracy and made further edits. Still, I could not have created such rigorous and engaging content without AI’s help. You can view one of many conversations here.

Second, I used it to create rubrics and personalized feedback for every student. For example, students read a post about retirement savings; I asked GPT4 to generate questions based on the text and to create a rubric. I used student first names to personalize their feedback, which I then cut and pasted into Canvas (enough cutting and pasting to justify buying custom keys). Here is a consolidated post with a fictional student name and sample feedback. Notice that I asked GPT4 to customize the feedback for a 7th grader. There’s still plenty of professional judgment to exercise before sharing this feedback with the student. However, these drafts free up my cognitive power to think more about using the information formatively and fine-tuning summative assessments. My subscription lets me get personalized feedback for up to 25 students every three hours.

Third, I created a reflection assignment with this prompt:

Create a set of short-answer questions … based on this prompt: ‘Review the feedback on … your assignments in Canvas. Choose your assignment with the most detailed and extensive feedback. Answer the following questions …:’ Use the most effective research-based techniques to create questions to build students’ metacognition. Write the questions at the 7th grade level.

After editing the 13 questions GPT produced and adding one of my own, my advice is to start more slowly and phase in reflections on feedback. The feedback assignment was largely successful, but the long list of questions intimidated some students. Next year, we’ll space out metacognition.

The possibilities for using AI to improve learning are vast. I used it in advisory to help students craft their vision statements for 8th grade. Perhaps I’ll ask it to help with my vision statement for teaching 12th grade in 2023-24. (I might need to use the beta version that searches current websites for help.)

The bottom line remains: What can we do to create the best readily achievable challenges to help our students learn content, practice skills, and get feedback to live their most purposeful lives? Using AI is now a critical part of that journey.

‘Choose Your Own Adventure’ Reading

A marketing and entrepreneurship teacher at Mount Miguel High School in San Diego, Mick McMurray is the creator of Grobots.ai:

Something unexpected happened when I started having students use ChatGPT as an assistant for their assignments—it replaced many of the books, websites, and articles I usually use.

For context, this was a “Shark Tank” project in a high school marketing class. Students were working in teams to develop a business pitch to investors. In place of the typical readings I would use, I created a series of ChatGPT prompts. For example, when it came time to decide on distribution channels, I provided students with this prompt:

“You are a business expert. We are collaborating.

I’m starting a new business. I already have an idea for a product.

Our goal for this chat is to decide on distribution channels for our marketing mix. We are figuring out the ‘place.’

First, you need to ask me about my product and target audience.

Once you know about my product:

1. Teach me about ‘place’ in the marketing mix.

2. Help me decide on the best ‘place’ for my product.

Ask questions as needed to get more information. You should be guiding me to make my own decision.”

Students cut and paste this prompt into ChatGPT. (Note: if you are using the paid ChatGPT, you can now share prompts so you can just send a link.) This prompt would lead ChatGPT to first ask for more information:

“Sure, I’d be happy to help you with your new business and guide you through the process of deciding on the distribution channels for your marketing mix. To get started, please tell me about your product and your target audience.”

This response was the start of a “choose your own adventure” reading for a student.

I settled on a four-step process for these “readings”:

First, students would have an individual chat with ChatGPT.
Second, they would engage in a structured conversation with their teammates (e.g., Think, Pair, Share).
Next, they would continue the chat on one student’s device to get to a group outcome.
Last, there was a short individual reflection specific to the task. In addition to providing an opportunity for metacognition, I used these for formative assessment.

As I watched students’ high levels of engagement with this process, I was struck with the observation that this was the application of Self-Determination Theory to readings. According to the theory, individuals have three basic psychological needs that, when satisfied, facilitate intrinsic motivation: autonomy, mastery, and relatedness. Interacting with an AI provides the freedom to explore and engage with the concept at one’s own pace and in a personalized manner. The AI can present information, answer questions, and provide feedback, allowing learners to actively participate and gain a sense of mastery. In this case, relatedness was created through the team component of the project, but future AI systems will undoubtedly offer more personalized and empathetic responses, building a sense of connection.

The autonomy of these chats was an easy-to-implement application of Universal Design for Living guidelines. I first had to help students get past their desire to get the perceived “right” answer and continue the chat to increase their understanding. They learned to make requests of ChatGPT, like “use simpler words” or “be more concise.” With guidance, students were able to make personal decisions about differentiation.

My greatest concern as we progressed through this lesson was that students did not have to think for themselves. This has been a common concern as educators have considered the impact of ChatGPT and AI on learning. While this may indeed be true when used to create a single output like an essay, my concerns dissipated once I saw the final presentations. From the quality of their final presentations, it became clear that ChatGPT served as a guide, not a crutch. The students’ creativity and critical thinking shone through, proving that AI tools can be valuable allies in the learning process

Thanks to Mary Beth, Paul, and Mick for sharing their thoughts.

Consider contributing a question to be answered in a future post. You can send one to me at lferlazzo@epe.org. When you send it in, let me know if I can use your real name if it’s selected or if you’d prefer remaining anonymous and have a pseudonym in mind.

You can also contact me on Twitter at @Larryferlazzo.

Education Week has published a collection of posts from this blog, along with new material, in an e-book form. It’s titled Classroom Management Q&As: Expert Strategies for Teaching.

Just a reminder; you can subscribe and receive updates from this blog via email (The RSS feed for this blog, and for all Ed Week articles, has been changed by the new redesign—new ones are not yet available). And if you missed any of the highlights from the first 11 years of this blog, you can see a categorized list below.

I am also creating a Twitter list including all contributors to this column.

The opinions expressed in Classroom Q&A With Larry Ferlazzo are strictly those of the author(s) and do not reflect the opinions or endorsement of Editorial Projects in Education, or any of its publications.