AI models now offer tutoring modes, like Gemini Learning Coach and ChatGPT Study and Learn. But are they any good?
We tried out Gemini’s Learning Coach to see how the model responds to learners’ requests. TLDR: Asking for a learning process will likely help students build historical thinking skills more than asking for coaching.
Using the What Causes Revolutions lesson from OER Project: World History 1750 I tested two learning scenarios.
Scenario 1: First, I told Gemini Learning Coach “I am a student in world history and I’m supposed to make a causal map of the causes of revolutions. How do I do that?”
Gemini Learning Coach gave me an introduction (see below) and linked to a video from YouTube called “Drawing Causal Diagrams.” A cool feature, but the video wasn’t specific to world history even though I’d provided that information. Our recommendation: tell students to let Learning Coach know they are using OER Project.
Once I shared that context, Learning Coach adapted, and moved on to identifying Key Causal Factors.
First, Learning Coach explained long-, intermediate-, and short-term causes then made an action plan for me. This is where the learning experience broke down. The action plan here is solid, but in execution Learning Coach is doing what we’d want the students to do. Instead of letting me choose the revolution (step 1) and brainstorm causes (step 2), the AI generated those for me then asked me to categorize them (step 3). The difference between tutoring and coaching is that coaching draws out a learner’s thinking through strategic questioning, and this example shortcuts that process.
Ultimately, Learning Coach guided me to map the relationships between nodes to create a causal map. This was a little tricky to do only using text. This is one reason we say that all AI tools have a limit to what they can do in a chat session. Finally, Learning Coach summarized how to build a casual map and offered me a multiple-choice quiz. If my goal was to have students practice their historical thinking skills, I couldn’t recommend this prompt exactly. Instead, I would want students to do the same thing that I did – Go through the process and assess what Learning Coach is and is not doing for them, then determine the accuracy of the information it provides against what they’ve learned in class.
Scenario 2: I tried to see if Gemini’s Learning Coach could help students with a peer review protocol. This time I asked for a study guide that a group of students could use to prepare to write a DBQ, and linked the specific task from OER Project: https://www.oerproject.com/World-History-1750/Unit-2/An-Age-of-Revolution?share=link#guid98e347dd-d4a1-4414-a89d-a5eb49725d41
The agenda above shows the whole Study Guide and Learning Coach’s directions for each step. The guidance in each step of the Study Guide seems to be more helpful than the first prompt with causal mapping above.
If I were teaching, I would extend the process by asking students to use their assignments from the unit to gather supporting evidence as part of their review. I’d also require that students put things in their own words, even the tempting “hint” from OER Project above.
In step 3, Learning Coach again goes a little farther than I would like in sharing information before students have had a chance to think through it themselves. But in a study group format, one student could be the facilitator who sets up the comparison matrix for the group but does not show the information that Learning Coach has already filled in. And the other students could add supporting ideas and evidence.
Finally, Learning Coach suggests 5 minutes for a thesis huddle. Again, there is an example that might steal students’ thunder or make it harder for them to form their own ideas. One idea is to put a notecard over the example thesis until they have written one together. Then they can compare their own ideas to the thesis from Learning Coach.
These interactions make me wonder if we should explicitly teach students to tell Learning Coach “I want to try this on my own. Don’t tell me the answers or show me an example until I’ve shared my thinking.”
Do I wish Learning Coach held back a bit? Yes. Is it a valuable tool for teachers and students? I think so.
Now, this was an unscientific experiment with lots of variables at play, so it would be interesting to see different results from others. What are some ways you might use Learning Coach? How might you guide your students to use this tool?
