AI to Support Learning
Now, I have a feeling that this blog is going to be a hard sell for some. When we talk about AI in secondary classrooms, we most often aren’t talking about the benefits. This is absolutely understandable because right now, the possible benefits are being swallowed up by students using generative AI to skip learning entirely. They are simply prompting ChatGPT to write their essay and then turning it in without a clue of what they were supposed to actually get better at by engaging in that process.
I don’t want to get sidetracked, but that process alone should be enough to convince us that the way we typically assess students in schools is focusing them on the wrong thing. A student who uses ChatGPT to complete an assignment is indicating that their goal has nothing to do with learning. Rather, their goal is the grade, and we’re now facing down Goodhart’s Law, which states that when a measure becomes the target, it ceases to be a good measure. This is why, now so more than ever, we should be asking hard questions about our assessment and grading practices.
However, the goal of this isn’t to talk about our grading practices as a whole, but rather, to give you an idea of what it could look like to use AI as part of the actual learning and assessment process.
Let me lay out a classroom activity that uses controlled forms of AI to point students towards their learning and then provide individual support to help them get there.
Now, before I jump in, I want to say a couple of things. First, neither of these tools I’m about to mention require students to have an account or to put in any personal data. Right now, we need to be more vigilant than ever before in vetting the tools that are rapidly emerging with the AI boom. Second, with each of these tools, there is a step where I’m incorporating intentional human interaction. Our classrooms need to be always human centered, and our processes with AI should start and end by focusing on humans and the human elements of learning.
Now, that being said, let’s dive in.
Say that you teach a 6th grade science class. In this class, you’ve been talking about ecosystems and the impact that a warming climate could have on them. There’s been direct instruction, kids have processed collaboratively, etc., but you’re coming up on a step of wanting them to write a paragraph explaining how a specific animal would be impacted by a change to their ecosystem as a result of climate change.
Building Background Knowledge
To start with, students are going to need to make sure they fully understand the concept and to process through their ideas around what animal they could talk about. This is where SchoolAI comes in with their Spaces feature.
What this tool lets me do is to create a custom chatbot that operates within a set of rules that I establish for it. For this example, I was able to tell it to focus on helping students brainstorm how climate change could impact and animal’s ability to survive, including migration, breeding, food sources, and more. I was able to clarify the grade level and give some tips for how I wanted it to interact with students (ask questions, be friendly, etc.).
Screenshot of a chat with SchoolAI’s custom chatbot, Sidekick.
The cool part is that I can also set up rules about what it will and won’t do. For example, I told it to do this if a students asks it to write something for it…
Screenshot of a chat with SchoolAI’s custom chatbot, Sidekick.
This allows us to reinforce that, while AI can automate some tasks for us, we still are responsible for engaging in learning.
Now, once we’ve built some background knowledge and scaffolded students to be able to come up with their example and brainstorm about it, it’s important to pause here and have students share what they’ve learned, both with a partner and with the teacher. Part of this is to reinforce the human element of the classroom, but it also helps clear up misconceptions or possible incorrect information.
Writing the Paragraph…with Some Support
It’s time for them to jump into the writing. In an ideal world, students would be able to submit a draft of everything, and they would get immediate feedback from the teacher. First off, we don’t have that kind of time, and secondly, we still can’t do it immediately.
However, AI can be a support here. For this step, I’d like to introduce Curipod. Not only does Curipod use AI to create interactive slide decks and presentations around any topic you want, but it also has a nifty little feedback feature built in.
One of their pre-made templates is called “write, feedback, repeat.” What this does is let students write a draft, receive AI-generated feedback, and then write a second draft all in one sitting.
A picture of a student device on the left receiving feedback and the teacher dashboard on the right.
It' is absolutely crucial here that we don’t just move on with the AI—generated feedback without pausing to reflect. I like the prompt, “Turn to a neighbor and explain what was helpful with this feedback and what you wish you’d gotten more of. Partners, if you have ideas, try to help them with what they wanted more of.”
I recommend this because not only does it allow us time to check those conversations to ensure they received good feedback, but it also (again) reinforces the fact that learning is a social endeavor. (As a note, I only have the free version of Curipod, but if you have the premium version, you can provide guidance for the feedback you want students to receive by suggestions skills or concepts to focus on.)
Once students have process their feedback, Curipod will give them another chance to write a second draft of their answer. The piece I like about this is that, for the most part, Curipod does a solid job of identify a strength from their answer. This means that going into their second draft they have evidence that they can be successful if they build on what they already did previously. This is huge in helping to boost student motivation to persist to more difficult levels of learning.
Expanding Even More with AI
If I wanted to take this a step further, I might use Brisk Teaching to hop onto the final Google Doc where I could give additional feedback. The nice thing about Brisk is that it gives me more options to customize the feedback generated by AI. I can explain what I want it to focus on, or I can also add in the rubric itself to focus the feedback.
As with everything, when students receive this feedback, I typically try to either sit down with them in writing conferences to reflect and evaluate the feedback, or I will have them do that with a partner.
Again, we always end with the human element.
Final Thoughts
AI is new, and often with new comes fear. Add that to the fact that many educators are facing growing frustration at the ways AI is being misused by students in academic contexts, and it can feel like nothing good can come of this.
If you know my work and my approach to life, then it shouldn’t surprise you that I was a skeptic who fell into this camp at first. Honestly, I was daydreaming about running away somewhere with no technology and just living in the woods (which, in all fairness, is basically always my dream, so I can’t blame ChatGPT for that totally).
However, as I learn more and more about how all of this functions and the different tools we have available to us as teachers, that fear and skepticism is starting to be a bit more balanced with hope, excitement, and optimism about what the effective and ethical use of this technology in the classroom might mean.
AI Tools
Tools referenced in this post to look into:
Additional tools worth exploring to see what GenAI has to offer:
