
AI has opened a world of new opportunities for student learning. I have discussed how instructors can easily add interactions to their learning material, tutors to their classroom, and digital laboratory simulations that allow students to test how a process changes with different inputs. But these are just the beginning, and new ways to improve student learning are emerging daily. Here are a few options to try in your classes.
Role-play
AI tutors do a remarkable job of providing students with Socratic dialogues that mimic what they would get from a human being, allowing them to learn course material and concepts anywhere and at any time. This success has led teachers to create AI-driven role-play scenarios for students to practice the types of interactions they will encounter in their careers.
For instance, a marketing instructor may currently have students create marketing campaigns for a hypothetical business. But a marketer also needs to sell the campaigns to their clients, which means answering questions and dealing with objections. They will likely need to revise their campaigns in light of these interactions.
AI can provide students with experience in this process. Perplexity is a good system for creating simulations as an instructor can create a simulation as a Space, allowing students to engage with it individually without seeing each other’s chats. A marketing instructor could tell the system to play a cautious client who is listening to a marketing campaign pitch. The instructor could also tell the system to include common questions that clients have, such as whether the campaign can be done more cheaply or why a certain media mix was chosen.
Further, the instructor might require students to upload their campaign materials as files and then pitch the campaign to the system in the chat via audio. They would answer questions, agree to make changes, push back against objections, and so on. The instructor might also have them revise the campaign in light of the interaction and return to the system to see what it thinks of the new work. Perplexity will remember each conversation’s history as a unique Thread, which the instructor can view. Each student would then submit the final product along with an overview of the interaction, including a reflection on what the student did and why.
Pretty much any profession will involve interactions that AI can help model, and there are a lot of instructors building interactive personas in AI. There are even video-generating tools that create lifelike avatars that respond to user input in real time. A counseling student can practice working with a client using one of these tools, while a business student can practice working with a boss. I will provide a more detailed discussion of these tools in a future article.
Historical recreations
One of the more interesting projects in genetics is to bring back extinct animals, such as mammoths, via their DNA. Similarly, an interesting area of AI exploration is in recreating historical figures from information about them. Hayk Grigorian, a computer science undergrad at Muhlenberg College, is building an AI agent that emulates the speech and worldview of Victorian England. His TimeCapsuleLLM is trained on historical texts from 1800–1850 London without any modern knowledge.
The model is still under development but has already yielded some interesting results. Grigorian asked the system about the year 1834, and it replied as follows:
It was the year of our Lord 1834 and the streets of London were filled with protest and petition. The cause, as many recounted, was not bound in the way of private, but having taken up the same day in the day of Lord Palmerston, the public will receive a short statement of the difficulties under which the day of law has reached us.
Grigorian looked up the events in London that year and found that there was a major protest related to Lord Palmerston, which the system went on to accurately describe. He then asked the system to create a passage about telephones, which had not been invented yet, and so it was clueless.
I have long wondered what it would be like to hear people talking in 1776 Boston, and projects like Grigorian’s bring us closer to realizing that possibility. I can see students in an ancient philosophy course creating a Plato agent that interacts with users as Plato did with his students. They would need to evaluate the tool’s responses against their knowledge of Plato’s philosophy and style and make adjustments as needed—an interesting learning activity that would give them experience working with AI agents.
Find and fix
Error identification is a powerful teaching tool. I use it as a fun side activity in my classes through daily grammar puzzlers. I give students a passage with a writing mistake and ask them to identify it. For instance:
The Green Bay Packers used running backs Aaron Jones and Jamaal Williams and fullback Danny Vitale to force Oakland Raiders linebackers into coverage, which were mismatches every step of the way. The trio found themselves wide open for Rodgers on short and intermediary routes, and Jones and Vitale were able to turn some of them into big plays with runs after the catch.
Can you spot the error? Intermediary should be intermediate.
Until now, I’ve had to either take these puzzlers from news articles or invent them, but AI makes it easy to generate them. I gave Google Gemini the prompt “Create a paragraph written for a college student that uses one of the words incorrectly. After the paragraph, add a sentence that explains the mistake,” and it responded with the following:
While the syllabus for the advanced macroeconomics seminar seemed daunting, I decided to approach the semester with a sense of tenuous enthusiasm. I spent my entire weekend at the library, meticulously annotating every chapter of the primary text to ensure I wouldn’t fall behind. Despite the heavy workload, I am confident that my dedication will eventually lead to a high GPA and a prestigious internship.
The Mistake
The word tenuous is used incorrectly here; it means weak, slender, or fragile, but the context of the paragraph describes a student who is being very firm, meticulous, and intense in their efforts. The word should be tenacious, as it is someone who refuses to give up.
Instructors can create similar find-and-fix examples related to their teaching subjects. For instance, a math teacher can use AI to generate a math problem solution with an error that students need to identify.
A variation on this method is to have an AI system produce multiple outputs in response to a request and have students evaluate them and pick the best option, explaining why they chose it. Returning to the marketing campaign example, an instructor might have students generate multiple marketing campaign ideas from a single AI system or ask three different systems to generate a campaign, then pick the best one and iterate on it. This is precisely how marketing professionals are using AI today, and thus the exercise prepares them for their future profession.
These are the kinds of opportunities for learning that are now available to teachers. The only limits are our own imaginations.
John Orlando is a professor of philosophy at Colorado Technical University and an educational consultant who leads workshops for faculty on how to use AI and other technologies to improve their teaching: jorlando2001@gmail.com.
The post New Opportunities for Student Learning appeared first on The Teaching Professor.
