ChatGPT isn’t the first AI language model, but it’s definitely the one that has taken AI mainstream. Beyond the media hype about cheating, there are some very practical uses for these new technologies for teachers.
This post covers six areas where teachers could use ChatGPT with examples of the kinds of prompt you can use to get the best results:
If you’re looking for a post on the basics of AI, language models, and ChatGPT, check this out first:
The Prompt is Key
Right now, one of the most powerful things you can learn about ChatGPT is how to write quality prompts. I put together a blog post in the “early days” of ChatGPT (a few weeks ago) on prompt engineering. In that post, I suggested 7 ways to craft better prompts:
- Be precise
- Avoid the “efficiency” trap
- Check the facts
- Iterate and improve
- Role play
- Remind, remind, remind
- Fill in the gaps
C: Clarity – Clearly define the task or intent of the prompt, including specific information about the output.
R: Relevant – info Provide relevant details, including specific keywords and facts, the tone, audience, format and structure.
E: Examples – Use examples in the prompt to provide context and direction for the output.
A: Avoid ambiguity – Focus on the key information and delete unnecessary details in the prompt.
T: Tinker – Test and refine the prompt through multiple iterations. Explore different input versions to discover the best results.
E: Evaluate – Continuously evaluate the output and adjust the prompt as needed to improve the quality.
For the examples later in this post, you’ll see that I’m using many of these techniques. Perhaps in future generations of these technologies the need to be so specific will be reduced. For now, I actually think that crafting a great prompt that outputs exactly what you’re looking for is part of the joy of working with this technology.
Remember, ChatGPT is not magic. Although the results are sometimes surprising, at the end of the day it’s just an algorithm doing its thing. The more accurate your prompt, the more likely it is the algorithm will select the correct words.
Practical Strategies for using ChatGPT in Education
ChatGPT can be a powerful tool for streamlining the lesson planning process, but it’s important to use it strategically. Using the technology to simply jump through an audit or school policy hoop becomes an arbitrary and unnecessary task. Instead, lesson plans should be written for a specific purpose, such as clarifying lesson goals and outcomes, fostering discussions about diversity and differentiation, and ensuring consistency with the curriculum and other teachers.
It’s important to use ChatGPT as a collaborative tool, rather than outsourcing the entire process to the model. This might involve using the model’s suggestions as a starting point for a team of teachers to workshop and refine. For example, instead of prompting ChatGPT to “generate a lesson plan for Year 7 English,” a team of English teachers could use the model to generate ideas for a specific theme or topic, and then work together to develop those ideas into a cohesive lesson plan.
Here are a few examples of how ChatGPT might become an additional member in your faculty planning team. I’d suggest using these prompts in a meeting and passing the ideas back and forth between the teachers in the room, and the model:
Lesson on <topic> should meet the following outcomes:
<copy/paste outcomes from curriculum documents>
Suggest learning intentions written in student-friendly language for an introductory lesson which covers <content>
This unit plan on <topic> requires four formative assessment checkpoints as part of a folio of work:
<copy/paste unit plan draft>
Suggest four formative assessment tasks and identify which week they should occur in the unit.
Suggest places where this lesson might bottle-neck or become less engaging, and suggest alternative approaches to improve the flow of the lesson:
<copy/paste lesson plan>
Generate a list of resources and materials we will need during this unit of work:
<copy/paste unit of work>
Role play: You are a teacher in our <subject> faculty. We are currently in a meeting to discuss a unit of work on <topic> for Year <year level> students. You are knowledgeable but highly critical and a little cynical. Your role is to critique and question the unit plan, and we will type our responses. Do not provide our responses, only your responses. Here is the unit plan:Note that the “do not provide our responses” is not always necessary, but it often produces the entire conversation if you do not include something like this.
<copy/paste unit plan>
We all have a lot of old materials lying around. While ChatGPT can competently create new materials (or help workshop new plans, as above), there’s no reason to discard all those resources we already created with human hands. In fact, using ChatGPT to update materials you produced yourself will give you insights into both the language model’s processes, and your own.
You could use ChatGPT to update old units of work to align them better with your values, and the values of your students. ChatGPT could take existing lesson plans, and redesign them to better fit your school’s teaching and learning processes and pedagogy. Or, you could take those tired old worksheets, drop them into ChatGPT, and use it to co-design a scenario based project or another more engaging lesson.
Here are a few example prompts you might use to refresh your existing materials:
<copy/paste a unit of work into ChatGPT>
Use this unit of work as the basis for a new unit outline. Keep the same core topics, but suggest more engaging activities and formative assessment methods. Focus particularly on adding group work and opportunities for students to engage in practical activities.
Explain the term gaps and silences in the context of critical literacy.
<copy/paste a lesson plan into ChatGPT>
Suggest gaps and silences in this lesson plan.
This is an example of out school teaching and learning framework:
<copy/paste T&L framework>
We also use the following approach for designing rubrics:
<copy/paste rubric design framework>
Update this existing rubric to better reflect our teaching and learning framework and school approach to building rubrics:
<copy/paste old rubric>
Here’s an example of the first prompt, showing a before and after of the first two weeks of a Pride and Prejudice unit plan. Use the slider to switch between the original and ChatGPT’s update.
In this earlier post I discussed going beyond just the “efficient” uses of ChatGPT and starting to look at how it could be used to improvise. or as an “analogy engine”.
Being able to improvise is definitely out of the realm of current AI language models. The machines can’t “think”, let alone think on the spot and in novel and interesting ways. We certainly can improvise, but it can be difficult to think flexibly in the middle of a 50-minute lesson during period 6 on a Friday, faced with a room of twenty-something tired adolescents.
So what if, during the lesson, we start using AI as an “improvisation aide”? For example, if a student is struggling with a particular concept, we might not have the energy to think up novel alternative explanations. But ChatGPT has no such problem, and it makes a great analogy-engine.
Picture the scene:
Teacher: <explains Non-Linear Equations>
Student: I don’t get it
Teacher: <explains it again, with a different example>
Student: <blank expression, shrugging. Yawns>
Teacher, to ChatGPT:
Generate three novel ways to explain Non-Linear Equations to a grade ten student at 2pm on a Friday
Imagine being able to generate on-the-fly essay topics for students who have a particular interest, or following a discussion of a text that has gone down a rabbit-hole. Or being able to quickly draw up scenarios and role plays for students to act out to help clarify their understanding of new concepts. AI has the potential to supercharge a teacher’s natural capacity to improvise and create engaging lessons. Here are a few more examples of how ChatGPT might be used on the fly during a lesson:
Generate three possible prompts for a personal response essay exploring the connections between <text> and <student’s favourite sport/hobby>
We are holding a class discussion on <text/topic>. We have started to discuss <whatever rabbit hole you’ve just gone down>. Generate a list of 10 discussion questions to help us explore these ideas further.
Generate 7 different scenarios for groups of three students based on <text/topic> and exploring <idea being discussed>. Provide descriptions of each student’s role in and the scenario in this format:
Student 1 role:
Student 2 role:
Student 3 role:
ChatGPT can be a great tool for personalising lesson resources and approaches. This can be as simple as adapting lesson materials to suit something the student is interested in (like the “personal response” prompt above), or as complex as designing an entire Individual Learning Plan.
One area that can easily be personalised is the complexity of the materials you present to students. I used to use “traffic light” resources – versions of the same article or text at three levels of reading complexity. The problem is that those resources are hard to find and time consuming to produce. This process is much easier with ChatGPT:
Reproduce this text at a Level 6 Flesch Kincaid grade level*. Keep the original tone and style, ideas, and structure:* I’m using the FK grade level here as it’s one of the most commonly used, but you could use other scales or simply ask for a year level.
^Make sure you’re not breaching any copyright by submitting a text to OpenAI!
This next one is a little more involved. I’ll provide an imaginary scenario, but one that will be familiar to many teachers.
Imagine you are a Year 7 homeroom teacher. You have recently attended some great professional development on working with autistic students (full disclosure – I’m on Reframing Autism’s board so I’m going to say it came from there!). The PD included reference to a paper on Universal Design for Learning (UDL) and inclusive education. You have an autistic student in your homeroom, and after some discussion with them you both agree it would be useful to take what you learned and add it to the student’s Individual Learner Profile for other teachers. This is not as easy as it sounds: you need to revisit the PD materials, break them down into a format that fits the ILP template including how it aligns with existing goals, and communicate this to the teachers.
You need a workflow:
1. Prime ChatGPT by checking its existing knowledge of UDL and the concept of ‘inclusive education’. This both contextualises your next prompts, and allows you to check that the output aligns with your understanding of these concepts.
What is Universal Design for Learning? What is inclusive education?
2. Add some specific context from the reading you enjoyed. Use select parts of the reading (the whole thing won’t fit) such as the abstract, a few key extracts from the introduction, and your favourite ideas from section on applying UDL principles. (Return “OK” is just what I add to get it to give a minimal response in between prompts)
Use the following extracts from a journal article as the basis for future prompts, along with the basic principles of UDL and inclusive education. Return “OK”:
3. Provide details about the ILP you’d like to update. Do not submit any identifying details about the student. Go through this part of the process with the student, using ChatGPT to suggest and refine goals.
STUDENT has an autism diagnosis. STUDENT’s Individual Learning Profile currently has three goals:
STUDENT would like to update their goals to reflect ideas from the materials on UDL and inclusive education. Using the above, create 3 goals from the student’s perspective to help them work with teachers in an inclusive education setting. The goals are written for the student, by the student, in student friendly language and the first person.
4. Use ChatGPT to help generate adjustments for this student that could be applied in most lessons.
Suggest five ways these UDL and inclusive education principles could be applied by STUDENT’s teachers.
5. Get specific. Now that ChatGPT has updated the student’s goals and the general adjustments to record in the ILP, it’s time to produce some tailored suggestions for different subjects.
STUDENT’s subjects are: English, Maths, Health and PE, Wood, Visual Arts, Music, Indonesian, Science, and Humanities. Using STUDENT’s goals, the five general modifications, and the information from the journal article, suggest three modification dot-points for each subject teacher.
6. Review everything with the student and their parent/guardian(s). Transfer the information into the ILP. Prepare to inform the staff of the updates.
My name is Leon. The teacher’s names are:
Health and PE: George
Visual Arts: Norman
Write emails to each of the teachers as follows:
Hi <teacher name>,
Just a quick update to STUDENT’s ILP. We’ve added some information based on a great PD around Universal Design for Learning, inclusive education, and autism. The goals and general modifications in the ILP have been updated, and I wanted to give you some specific suggestions for your subject area. Here’s a few ideas:
<suggested modifications for subjects>
Hope that helps!
So here’s the entire workflow from start to finish:
- Prime ChatGPT and check accuracy of existing knowledge
- Add specific context from relevant materials
- Introduce the ILP and update goals
- Suggest general modifications
- Suggest subject specific modifications
- Write emails to communicate to teachers
ChatGPT is not an original thinker, but you are. One of the best ways to use ChatGPT – as indicated by the examples throughout this post – is as a collaborator rather than a replacement for teachers.
There are lots of ways both teachers and students could collaborate with the model (assuming we get out of the murky swamp of banning, blocking, and obscure Terms and Conditions).
Some universities have already started to adjust policies, requiring students to cite which tools they have used in the production of assignments – tools including Grammarly and ChatGPT. This seems like a much more proactive response than simply blocking the website.
There’s a line somewhere between using ChatGPT in collaboration, and getting it to do all the work. In the first strategy, planning, I wrote about getting ChatGPT to assist with lesson plans, rather than generating them entirely. The same applies for other styles of writing, for example:
Write an outline for a <subject area> handout for students on <topic>. Use roman numerals and markdown to format headings and subheadings.Markdown is a lightweight text formatting code. It can make your ChatGPT output a bit more organised. Click here for more details.
These are the first two paragraphs for an essay on <text/topic>. Brainstorm 5 possible directions for the next paragraph.
If we ever see ChatGPT being permitted for use by students in secondary classrooms, then we might also be able to encourage prompts like these, allowing students to collaborate with an AI “teacher’s aide”:
This is the handout and questions provided for this lesson:
Produce a dot-point summary of the handout and explain each of the questions and how I should approach them.
These are my notes from a lesson on <topic>:
Turn them into an outline for a 3 minute oral presentation.
In Strategy 4: Personalising, I ended with an example of using ChatGPT to communicate with colleagues about a student’s ILP. One major issue facing teachers is the pressure of communicating quickly and effectively with students, parents and colleagues.
A single teacher with 6 classes might have over 170 students. It’s unlikely that you’d have to speak to all of those students’ parents or guardians at once, but it’s likely that at some point during the year you’ll have to contact a good number of them.
It can also be challenging to communicate clearly when there’s an emotional component to the message. It might be an email to a parent to organise a meeting after an explosive incident in a class. Or, maybe it’s an email to a colleague who has just crossed one of your boundaries.
ChatGPT can be used to help draft, rewrite, and take some of the sting out of written communication. Aside from emails, ChatGPT can also assist with writing other forms of school communication such as reports and newsletter articles.
Rewrite this email to make it neutral and not accusatory:
<copy/paste email draft>
Draft an email to parents of <class> to congratulate the students on their efforts on the recent field trip. Here is a newsletter article about the field trip for context:
<copy/paste newsletter article>
Use the following feedback from students to create a script for an assembly item reflecting on the recent visit from <incursion>
<copy/paste survey form responses>
Draft a report to the school community from the Director of Learning and Teaching which focuses on the following achievements this year:
<list of achievements>
Augmenting teachers, not replacing them
Hopefully you’ve seen that throughout this post the suggestions aim to augment teachers’ practices, not replace them. ChatGPT can’t think for itself, and in a future post I’ll discuss some of the very real ethical concerns with the technology as a whole.
If we don’t engage with the technology, then we’re doing our students a disservice. For all the hype surrounding AI in the media at the moment, this technology will drastically change our education and work lives.
In a few months, ChatGPT might have been knocked off its perch by another model. But getting your head around the potential uses of large language models in education will help you to prepare for whatever comes next.
Got something to say or a question about this post? Want to talk about school PD or a conversation around policies and processes for AI? Get in touch: