AI and Assessment: Helping Students Engage in the Learning
As an English teacher, the introduction of AI tools like ChatGPT into the mainstream of society brought with it plenty of headaches. Like many others, I initially turned to technology to solve problems brought about by technology. I relied heavily on AI detection software before I learned how frequently those tools produce false positive and negatives. I used revision history and Draftback to check to make sure kids weren’t copying and pasting, and then students showed me all the TikTok videos that taught them how to avoid that being detected.
I learned quickly that policing it after the learning had been demonstrated in a final product was going to turn my entire career into a game of whack-a-mole, and I wasn’t willing to let that happen.
To attempt to change that, I asked myself two things:
What do students actually want to engage in?
How can I make their process of learning more prominent?
For the first question, I started with the best place I could think of to start: asking the students. Their responses were incredibly insightful. When I asked, “What are some qualities of assignments you would actually want to do yourself instead of just letting AI do it?” students gave me a variety of answers that helped refocus me on assessments that were meaningful.
One student said, “I want to be able to get to choose what I do.” When I pressed on that a bit, they clarified that it’s not that they want to just do anything, but if they’re going to write, they want to be able to write about something they’re interested in. They want to connect with their passions and interests, and they want the autonomy to pursue those within our curriculum to help make it meaningful.
Another student said (and this one doesn’t surprise me) that if they feel like they can do it well on their own, they won’t use ChatGPT. Though, to be fair, when I pushed on this one, they were pretty transparent that if they don’t see the assignment as relevant or helpful, they’ll still use ChatGPT. This one taps into a much deeper concept around motivation where students must have a certain level of self-efficacy (evidence of previous success to justify effort toward future success) to be willing to engage in the task.
I want to add one more because I think it’s a key one. A couple of students said that if the teacher has talked about generative AI tools in their class and the students know they are aware of them, they are less likely to use tools like ChatGPT to complete the assignments. While it’s really tempting to try to bury our heads in the sand and pretend it’s going away, it won’t. It’s here to stay and will have an impact on our classrooms. It’s key that we begin working to understand what it is, what it means for our classrooms, and what it means for our students’ futures.
With everything students shared, it all came back to the ideas of what helps build intrinsic motivation: autonomy, mastery, and purpose. Students want the autonomy to choose things that connect to what matters to them. They want the ability to pursue mastery and know that they can achieve success. Lastly, they want to know that what they are doing is building towards bigger picture.
If we can be intentional in our planning and developing learning experiences that tap into these elements, it is more likely students will be willing to engage in them personally instead of under false pretenses by using ChatGPT to skip the hard work of learning.
Now, as much as I would love to say that this is a perfect fix, we all know it isn’t. We could make the most engagement assignment in the world, and students will still try to use generative AI to skip past it and get back to playing the game they love, scrolling TikTok, or whatever other activity is more interesting to them than school.
As such, we can’t rely solely on their intrinsic motivation to keep them from skipping learning. We also have to ensure we have a process to capture the learning by focusing on the process of it instead of the product.
Here are a few ways that I’ve worked to emphasize the process of learning over the product.
1) Have students complete a learning memo in connection to their products.
This can be exceptionally simple, but it can still have a major impact. In brief, a learning memo asks students to explicitly talk about how they demonstrated the expected learning outcomes in their product. If I’m asking students to integrate evidence into their writing, I might have them pull out an example and explain what they did and why it was effective. Notice that even if the student used AI tools to complete the assignment, this still requires a certain level of thinking and processing to support the learning. This can be done in a written format or (my favorite) a screencast recording of them explaining their thinking while going through their work.
2) Engage in learning conferences with students.
It may happen someday, but at least as of now, generative AI doesn’t have vocal cord implants yet. This means that a conversation with a student is one of the most foolproof methods we have of gathering accurate information about student learning in the age of AI. While my forthcoming book dives into this in much more detail, these really don’t need to be complication. It can be as simple as sitting down with a student and asking, “What do you know about the concept we’re studying right now?” Even a short conversation like this will elicit a surprising amount of evidence of student learning. I often use the analogy that assessments are the bricks, but my learning conferences are the mortar that holds everything together.
3) Use a “my favorite mistakes” assignment.
This one came about organically with my students. They’d heard my say over and over about how valuable mistakes are in your learning, and we often would highlight a mistake and talk about what we could learn from it. From this, we developed an activity where students would record their mistakes throughout a learning process or unit and then explain what they learned from that mistake. This has actually turned out to be huge in helping with the assessment process with generative AI around.
The piece that I want to emphasize is that, while we can panic (and rightfully so for many of us), we can also look at the silver lining in this. Generative AI is exposing areas of our practice where we have an opportunity to really focus on learning more than we have in the past. It’s absolutely frustrating at times, I know, but I still have hope that there’s the potential to move us somewhere positive.
If you like this concept and want to dive in more, I have a book coming out called Hacking Student Motivation that dives into meaningful changes we can make to our assessment practices to remove some of the barriers to motivation students often feel in school. Click here to join the mailing list to stay up to date about the book release, exclusive perks, and an assessment mastermind group that meets once per month.
