Introduction to Active Learning
Active learning is an instructional approach that involves engaging students directly in the learning process through activities and discussions, as opposed to passive reception of information. Active learning is a strategy to help students expand their thinking about content beyond the immediate context provided in traditional fact memorization and process replication exercises. Active learning involves “hands-on” and “minds-on” activities that engage students through interactive discussions and activities (Yannier et al., 2021). Active learning includes gamified learning, which is built on interactive elements to make learning more engaging and participatory (Lampropoulos & Sidiropoulos, 2024). Active learning encompasses practices and activities which promote student-centered learning (Lord et al., 2021).
The Effectiveness of Active Learning
Evidence strongly suggests that active learning is a superior instructional strategy. Lampropoulos and Sidiropoulos (2024) explore the impact of active learning through gamification on learning outcomes. The three-year longitudinal study shows that gamified learning significantly improves student engagement, motivation, and academic performance, particularly in success and excellence rates when compared to both traditional and online learning settings. A meta-analysis by Lord et al. (2021) found that active learning methods are more effective than traditional lecturing in improving student achievement in humanities and social sciences. In addition to improved learning outcomes, asking students to engage with content in ways that are specific to their backgrounds and experiences can promote emotional regulation and identity formation (Gotleib et al., 2024).
A Practical Example: Learning to Ride a Bicycle
A quick example: Most of us probably learned how to ride a bicycle the “traditional” way in which you mounted the bike (with assistance from a more experienced bike rider if you were lucky) and continued to attempt an upright / rubber side down travel path until you fell or crashed. It’s unlikely that any of us learned to ride a bicycle by reading a book or listening to a lecture. However bruised and shaken, we learned through individual and self-directed practice. The recent rise of the balance bike (Wikipedia, 2024) illuminated a way to guide the active learning in a way that is more proficient and less painful. Personally, as an instructor who implemented a balance bike strategy (just a lowered seat without training wheels, not an actual balance bike) I was blown away by the instant progress made when balance and thrust are taught sequentially instead of simultaneously. In short, active learning guided by experts is even better than passive learning or active learning without the helpful hand of expertise.
In pursuit of active learning strategies which allow students to contextualize the content to their background knowledge, personal experiences, and individual perspectives while guiding them through the learning process we’ll explore four ways that AI can serve as a teaching tool for your subject area.
Four Ways AI Can Serve as a Teaching Tool
1. Building Personas to Explore Possibilities:
Introduce students to someone who doesn’t exist: Platforms like ChatGPT can create personas based on specific instructions you provide or material you reference. Students can interact with these personas to propose an argument, query for possible opinions, and analyze how well the persona represents the intended simulation.
Example Activity 1 (In-Person Class): Tell ChatGPT to answer subsequent questions as an 18th century privateer operating in the Caribbean seas. Let students ask questions about specific privateering techniques, habits, or pro-tips. After the persona has generated a good amount of responses, have the class discuss the possible accuracy of the claims made. Furthermore, guide a discussion about what assumptions were made and how assumptions are used in research and exploration.
Example Activity 2 (Online Class): Supply the class with a set of opinions written by a Supreme Court judge who no longer exists. Ask students to read those opinions, and then supply the class with a proposed legal challenge. Have the students predict how the judge would rule based on past verdicts rendered. Finally, have the students give the same opinions to Claude / ChatGPT and ask the LLM to adopt the persona of the judge who authored those opinions. Once the persona is created, students will pose the legal challenge to the LLM and ask for a verdict. Have students report on whether the verdict matched their prediction and what confounding cultural or legal activity has happened since those opinions were written which might impact that verdict.
2. Imagining Personas to Simulate Debate and Practice Teaching:
Have the AI act as a younger / less experienced student. AI platforms can simulate a specific level of understanding or background and respond from that position. Students can work to explain a concept or position to the simulation. Each conversation will take a different turn, and students can practice refining arguments or rephrasing explanations.
Example Activity 1 (In-Person Class): Prepare the AI with a position or level of understanding on a topic, and have the class convince the simulation of a new idea or perspective using concepts from prior class sessions or class preparation. Prepare the simulation as a sophomore level university student with a Newtonian understanding of physics and have your students teach the AI modern physics. Specify that they must use concepts taught in your course alongside any other tactics which are necessary. You might prepare the AI to respond more favorably to specific arguments (see: prompt engineering) from resources you’ve already provided your students.
Example Activity 2 (Online Class): Provide a prompt which prepares the AI to adopt a position or level of understanding on a topic. Ask your students to educate / convince the LLM of the correct / updated understanding and report their process, the issues they encountered, and what line of thinking ultimately convinced the AI. In the example below, I direct the AI to adopt the position of a middle-school student with a specific idea of mathematics, and then I attempt to convince it of a different position (note that my prompt has a typo which the AI was able to interpret correctly):
|
👨 [Me] |
| 🤖 [AI Persona - 7th Grade Student] Sure, I can be a 7th grade student. Let's start the discussion! |
| 🤖 [AI Persona - 7th Grade Student] Hey, I think math was invented by people. It's just a tool we made up to help us understand the world. Why do you think math is discovered instead? |
| 👨 [Me] Do you agree that math "works"? |
| 🤖 [AI Persona - 7th Grade Student] Yeah, I agree that math works. It's really useful for solving problems and figuring things out, like in science and engineering. But just because it works doesn't mean it wasn't invented, right? Can't something be invented and still work really well? |
Between these screenshots ensued a longer conversation in which I provided analogies and perspectives in an attempt to change the persona’s thinking. The persona continued to ask questions and reframe my position, and forced me to narrow the conversation to a few specific examples.
| 🤖 [AI Persona - 7th Grade Student] Yeah, that is pretty cool. I get what you're saying about discovering math like finding clues. But couldn't it also be that we just made up all these rules, and then we keep adding new rules that fit with the old ones? Like, if we made up addition and then subtraction, maybe we made up multiplication to fit with them. And then, when we see how the planets move, we just use our made-up rules to describe what we see. How do we know we didn't just invent a really good way to explain things? |
| 👨 [Me] That's a really good point. What seems to support discovery is that in all of the math applications, we haven't found an instance in which we've had to invent a new mathematical system to explain or account for something. Whether the subject is biology, finance, astronomy, engineering, or marketing. |
| 🤖 [AI Persona - 7th Grade Student] I see what you mean. So, you're saying that if math was just invented, we might have to come up with new systems or rules when we find new things, but we haven't had to do that. Instead, the same math we've always used works for everything we discover. That does make it seem like we're uncovering something that was already there. But how can we be sure that we won't ever need to invent new math in the future? |
The persona finally agreed, but the entire conversation was a bit long for this post. Each conversation will be different for each student, so comparing notes and / or screenshots of some of the more interesting responses might be a good follow-up conversation.
3. Explain to me like I’m (insert characteristic here).
Ask the AI to explain a single concept in multiple ways. We’ve all had to shift our approach and vary our examples for students. AI can allow students to simultaneously work to describe (and reflect on) their background and perspective while thinking about difficult concepts.
Example 1 (In-Person Class): Have students share (before class) one idea they are struggling with in the class. Choose one or two which are most common, and open class asking students to describe themselves using personal experiences and interests. You might have several who have an interest in hospitality or a few who all like to travel. Let the class help phrase an identity and have the AI explain the concept using those parameters. Your prompt could be “Explain the difference between Attention Deficit Hyperactivity Disorder and Oppositional Defiant Disorder like I’m a hospitality professional. The response from ChatGPT explains guest behavior in categories such as inattention, hyperactivity, impulsivity, etc. as well as how someone in hospitality services might support guests exhibiting certain behaviors. You can have students evaluate the AI response(s) and reflect on their own learning with you.
Example 2 (Online Class): Have students create several AI personas to explain a difficult concept. They can report the prompt / persona, reflect on how this helped them learn, and evaluate the responses in relation to the explanation they received. In this example I ask the AI to explain a concept to three different audiences.
First, I prompted ChatGPT to explain statistical P values to me like I’m a professor of engineering:
| 👨 [Me] Explain statistical P values to me like I'm a professor of engineering. |
|
🤖 [AI Persona - Engineering Professor] Sure, let's dive into P values from the perspective of an engineering professor. What is a P Value? A P value (probability value) is a statistical metric used to evaluate the strength of the evidence against a null hypothesis. It quantifies the probability of observing your data, or something more extreme, assuming that the null hypothesis is true. Null Hypothesis and Alternative Hypothesis
Interpreting the P Value
|
This response was the most robust.
Next, I asked ChatGPT to explain statistical P values to me like I’m a high school track athlete 🏃:
| 🤖 [AI Persona - High School Track Athlete]
Interpreting the P Value
The significance level, often set at 0.05, is like a cutoff point. If your P value is below 0.05, you can be pretty sure the change is real. If it's above 0.05, you're not so sure. P Value in Hypothesis Testing 1. Set up hypotheses:
2. Choose significance level (α): Usually 0.05. 3. Collect data and perform a test: Maybe you compare your race times before and after the new training. 4. Calculate the P value: This tells you how likely it is that your results are just random. |
This response contextualized the content for evaluating the efficacy of a training program, which could be helpful to a student who is already interested in this topic area.
Next, I asked ChatGPT to explain statistical P values to me like I’m a 5th grader:
|
🤖 [AI Persona - 5th Grader] Alright, let's explain P values in a way a 5th grader can understand. What is a P Value? Imagine you're flipping a coin and you want to see if it's a fair coin (one that lands on heads or tails equally). A P value is a way to help you decide if what you see is just luck or if something unusual is going on. Null Hypothesis and Alternative Hypothesis
Interpreting the P Value
|
The response probably made some cultural assumptions, but the focus was around coins. I’m not sure how many elementary students use coins, but the effort was nice. I could prompt it to use dice… do kids play dice games?
Finally, and just to see what happened, I asked ChatGPT to explain statistical P values to me like I’m a Portuguese explorer in the early 18th century:
|
🤖 [AI Persona - 18th Century Portuguese Explorer] Ah, esteemed Portuguese explorer of the early 18th century, allow me to explain the concept of P values in a manner befitting your adventurous spirit and keen intellect. What is a P Value? Imagine, good sir, that you are charting unknown waters and you wish to determine if a newly discovered route to the Indies is indeed shorter than the well-trodden paths. A P value is akin to a navigational tool that helps you ascertain whether your findings are due to mere chance or if they hold genuine significance. Null Hypothesis and Alternative Hypothesis
|
I’m not sure about the language (why not in Portuguese?), but the contextualization around sailing is not bad. I should have specified explaining the subject matter to an explorer who was worried about sailing off the edge of the flat Earth.
4. Play it again
Have students write song lyrics about a topic which is foundational to your course. Use one of the song generation AI platforms to produce a song of their favorite style. Even if students use an AI to generate the song lyrics, the final product (if accurate) should be memorable for the student. To this day I still sing the ABC song when considering letter order. Or is that the Twinkle, Twinkle, Little Star song?
Example Activity 1 (In-Person Class): Have students share their lyrics and songs. Choose the best stanzas and phrases and create an even more accurate song choosing more precise language and order.
Example Activity 2 (Online Class): Students can share their final product through discussion board links and provide peer feedback. This might be on everything from technical accuracy to preferred musicality as long as they are listening to the concept and thinking about the material. I’ve linked to two versions of a song created by Suno.com about production curves and price below (your students will likely come up with much better results).
Conclusion:
By leveraging AI tools, faculty can promote personalized, interactive learning experiences that keep students interested in the traditionally difficult areas of our disciplines. Embracing AI not only improves the efficiency of teaching but also prepares students for a future where AI plays an increasingly significant role. As the saying goes, “If it feels like work, people do less of it. If it feels like play, people do more of it.” I’m hopeful these suggestions are interesting enough that your students do more of it.
References
- Gotlieb, R. J. M., Yang, X.-F., & Immordino-Yang, M. H. (2024). Diverse adolescents’ transcendent thinking predicts young adult psychosocial outcomes via brain network development. Scientific Reports, 14(1). https://doi.org/10.1038/s41598-024-56800-0
- Lampropoulos, G., & Sidiropoulos, A. (2024). Impact of gamification on students’ learning outcomes and academic performance: A longitudinal study comparing online, traditional, and gamified learning. Education Sciences, 14(4), 367. https://doi.org/10.3390/educsci14040367
- Lord, T., Prince, M., Stefanou, C., Stolk, J., & Chen, J. C. (2021). Effect of active learning versus traditional lecturing on the learning achievement of college students in humanities and social sciences: a meta-analysis. Higher Education. https://link.springer.com/article/10.1007/s10734-021-00775-0
- OpenAI. (2024). ChatGPT (Mar 18 version) [Large language model]. https://chat.openai.com
- Yannier, N., Hudson, S. E., Koedinger, K. R., Hirsh-Pasek, K., Golinkoff, R. M., Munakata, Y., Doebel, S., Schwartz, D. L., Deslauriers, L., McCarty, L., Callaghan, K., Theobald, E. J., Freeman, S., Cooper, K. M., & Brownell, S. E. (2021). Active learning: “Hands-on” meets “minds-on”. Science, 374(6563), 26. https://doi.org/10.1126/science.abj9957
- Wikipedia contributors. (n.d.). Balance bike. In Wikipedia, The Free Encyclopedia. Retrieved July 10, 2024, from https://en.wikipedia.org/wiki/Balance_bike