Using AI to make teaching easier & more impactful
Here are five strategies and prompts that work for GPT-3.5 & GPT-4
I don’t need to start this post with the usual predictions that AI will transform our classrooms. It is obviously happening. Students are cheating with AI. Students are getting help with AI. I have required AI for all my classes this semester, and I hear more instructors are doing the same. GPT-4 tutors are being launched by large organizations (both Khan Academy and Duolingo currently have them). The world is changing fast.
But one thing that is not changing is the best way for people to learn. We have made large advances in recent years in understanding pedagogy - the science of learning. We know some of the most effective techniques for making sure material sticks and that it can be retrieved and used when needed most.
Unfortunately, many of these advanced pedagogical techniques are time-consuming to prepare, and many instructors are often overworked and do not have the resources and time to add them to their teaching repertoire. But AI can help. In the rush to deliver AI benefits directly to students, the role of teachers is often overlooked. AI tutors, as exciting as they are, do not replace the complex role of a teacher in front of a class. But not enough effort seems to be going towards applying AI to help instructors. We have a new paper that tries to remedy that gap, by providing some research-backed approaches to pedagogy, and the AI prompts (for GPT-4, GPT-3.5, and other AIs) to implement them. You can read the paper here, but I wanted to summarize some of those approaches.
Strategy 1: AI-created examples
Explaining complicated ideas often means giving students lots of examples so they can truly grasp what you mean. When teachers provide multiple examples, it helps students understand abstract concepts in a real-world way, challenges them to think critically, and shows how subtle aspects of ideas can work in different situations. Ultimately, this approach makes it easier for students to apply what they've learned in new areas.
However, coming up with good examples is hard work for teachers, who already have enough on their plates. They have to think about whether examples will capture students' interest, connect to what they're learning, and have the right amount of detail without being too simple or too complex. The balance is key, since examples that are too complicated will just confuse students, while ones that are oversimplified won't actually teach them much.
AI can help teachers generate a ton of diverse examples tailored to what students care about and what they're working on. With AI-generated examples, teachers can make sure they're accessible but also informative. Of course, the teacher remains critical, because they need to vet the ideas and decide how to deploy them. AI just makes the job easier.
You can click this link to get Bing to generate examples: https://sl.bing.net/bePdl4o9xf2 (this new feature of sharing links to in-process conversations is a little like sharing programs - paste it into a browser with Bing access to start the prompt)
Or you can use ChatGPT, and paste this prompt in: I would like you to act as an example generator for students. When confronted with new and complex concepts, adding many and varied examples helps students better understand those concepts. I would like you to ask what concept I would like examples of, and what level of students I am teaching. You will provide me with four different and varied accurate examples of the concept in action.
In the paper, we discuss more about how to assess the quality of examples. You can share the ones you like directly with the class, or use them as part of an in-class exercise. For example, you can ask students to compare and contrast these examples: what different aspects of the concept does each highlight?
Strategy 2: AI-created explanations
Helping students truly understand new ideas is key to teaching them anything. To do this, teachers need to get where their students are coming from, carefully plan how they'll explain things, give students hints to follow along, and use analogies to bring big concepts down to earth. The goal is for students to be able to explain what they've learned in their own words.
But coming up with lots of ways to explain a single topic can take a ton of time and effort. It can be hard figuring out how to pitch ideas at just the right level, include background info, and adapt for different learning needs. This is where AI can help out.
AI can generate different explanations, walk through ideas step-by-step, and add in more examples. If some students are struggling, AI can provide simpler summaries to get them caught up. Teachers can also use AI to improve their own explanations by making them simpler or adding more examples. But it's important to remember AI-made explanations should be a starting point. Teachers need to check them over before using them with students.
Prompt for Bing: You can use the following link to get Bing to generate explanations: https://sl.bing.net/koA1v8uUzw4.
For ChatGPT, try: You generate clear, accurate examples for students of concepts. I want you to ask me two questions: what concept do I want explained, and what the audience is for the explanation. Provide a clear, multiple paragraph explanation of the concept using specific example and give me five analogies I can use to understand the concept in different ways.
We describe how to check these results for effectiveness and errors in the paper, but, once you have vetted these examples, you can use them in class, hand them out to students as a study guide to supplement their knowledge, or use them as the basis for exercises.
Strategy 3: Using AI to develop low-stakes tests
Students hate tests, but they are some of the most effective learning tools we have. Contrary to the popular belief that tests only serve to assess knowledge, they actually play a pivotal role in the learning process itself. By incorporating repeated testing and knowledge retrieval, students are better equipped to retain information in the long term.
Low-stakes tests offer active retrieval practice, encouraging students to recall information from memory, which in turn enhances their ability to remember and retrieve information later on. And these tests provide valuable feedback on students' understanding of the material, helping them identify gaps in their knowledge and adjust their learning strategies accordingly. This not only aids in information processing but also prepares students for high-stakes exams.
While low-stakes tests has clear benefits, they can be hard for teachers to put together. Creating good questions, providing scores and feedback, and ensuring questions match what students should know takes significant time and effort. AI can help. They can generate practice questions and provide targeted feedback, allowing instructors to focus on teaching instead of test-making. With AI support, students may learn concepts more deeply through continual practice that sticks with them beyond the lesson.
Here’s how to do it in ChatGPT: You are a quiz creator of highly diagnostic quizzes. You will make good low-stakes tests and diagnostics. You will then ask me two questions. (1) First, what, specifically, should the quiz test. (2) Second, for which audience is the quiz. Once you have my answers you will construct several multiple choice questions to quiz the audience on that topic. The questions should be highly relevant and go beyond just facts. Multiple choice questions should include plausible, competitive alternate responses and should not include an "all of the above option." At the end of the quiz, you will provide an answer key and explain the right answer.
The expertise of the instructor in avoiding AI errors is key here, as is their judgement in the types of questions to ask. Once that is done, there are many ways to use these quizzes. One option for including tests in a class is integrating test questions in a discussion or lecture. They give instructors insight into any misconceptions or errors students might have and allow them to make a decision: Is it time to move on, or do I need to adjust the lesson?
Low-stakes tests can be given to students in class as a group exercise in which teams report out their responses, followed by a class discussion. They can be assigned as individual classwork or homework or posted in an online discussion forum. Additionally, tests can be distributed in class, and after completing the tests, students can then be given the answer key. They can compare their responses with the correct responses. A follow-up to such an exercise might be a reflection: What skills do you think you need to work on? How might you improve?.
Strategy 4: Assessing what students know, and what they are confused by
Short check-in exercises are key for helping students and teachers understand the course material. They provide real-time feedback, allowing them to identify gaps in knowledge and areas that need clarifying. While grading assignments shows students their progress, these informal checks for understanding promote active learning and help motivate students by demonstrating that the instructor genuinely values their needs.
One such assessment method, commonly known as the "1-minute paper" or "muddiest point" exercise, promotes active learning and self-reflection. Students are encouraged to summarize their knowledge, pinpoint any uncertainties, and share their perspectives on the material covered. This not only helps identify gaps to be addressed in subsequent classes, but also fosters increased student engagement and motivation by demonstrating that instructors are genuinely responsive to their needs.
To design such an assessment, instructors can choose a specific focus, such as an activity, topic, or class discussion. They can then formulate a question that will reveal both what students have comprehended and what they find perplexing. Potential questions might include: "What was the most important idea or concept discussed in today's class, and why is it significant?" or "What has been the most challenging concept so far, and what aspects did you struggle to understand?" By prompting students to think more deeply about the course material, these assessments can pave the way for a more enriching and engaging learning experience. Then they can use AI to help summarize the results.
To have the AI help quickly summarize student responses, instructors can create a Google Doc or any shared document and ask students to submit their responses. Then, instructors can submit a set of collective responses to the AI with the following prompt: I am a teacher who wants to understand what students found most important about my class and what they are confused by. Review these responses and identify common themes and patterns in student responses. Summarize responses and list the 3 key points students found most important about the class and 3 areas of confusion: [Insert material here]
You can obviously use this analysis after class to help you see where students stand, or even in-class to see what points you might want to cover (or ask students to cover) further. Students also often love these sorts of exercises because they feel like their feedback is being listened to.
Strategy 5: Distributed practice with AI
Incorporating distributed practice into the learning process is important for helping students build strong and adaptable knowledge. Unlike massed practice, where topics are taught one after the other without much connection, distributed practice means revisiting material several times over an extended period. This approach helps students form better mental models, remember those half-forgotten facts, and really get a deeper understanding of concepts.
Now, even though it's useful, fitting distributed practice into the mix can be a bit tricky because of how course materials are usually structured and students' preference for massed practice. But AI can help. One way to bring distributed practice into the classroom is by introducing a topic and reviewing it after specific intervals, like a week, a month, and at the end of the semester. Teachers can use AI to whip up brief topic overviews and questions for ongoing assignments or assessments, adjusting the difficulty level as needed.
Linking new ideas to concepts students already know promotes deeper learning, and AI can lend a hand by weaving past topics into lectures or discussions. By asking the AI to find relationships between concepts, teachers can present various connections between ideas, making them clearer for students.
You can use the following link to get Bing to generate distributed practice exercises and tests: https://sl.bing.net/hnKI78bzvzw or you can ask ChatGPT: You are an expert teacher who provides help with the concept of distributed practice. You will ask me to describe the current topic I am teaching and the past topic I want to include in distributed practice. You will also ask me the audience or grade level for the class. Then you will provide 4 ideas about how include the past topic into my current topic. You will also provide 2 questions I can ask the class to refresh their memory on the past topic.
Once the instructor has some good responses from the AI, distributed practice can be implemented in a number of ways. In terms of timing, scheduling exercises and tests that space out practice across a course and that allow for some forgetting, is optimal. For instance, once students show evidence of understanding about one topic, practice of that topic (in the form of an assignment or a quiz) may be scheduled once some time has passed; students will need to work hard to pull out that knowledge from memory and that effort will help them access this information next time. You can use the AI output to make sure that happens.
Teaching with AI help
AI is an incredibly exciting tool for teaching. It is widely-available, inexpensive, and fast to experiment with. We hope instructors will try some of the techniques in the paper to get a sense of how AI can make applying complex pedagogical approaches more effective and less burdensome.
And, for those worried about AI, it is worth remembering that, despite decades of hype from VCR classes to Massive Online Courses, technology has not replaced teaching. Instructors, and class interaction, play a vital role in making learning happen. AI allows new forms of learning and pedagogy that can benefit instructors, and their classes. As instructors, we have to experiment to learn the best techniques that work for us, and our students.
Thank you for continuing to be a voice for how AI can be used well. I appreciate your thoughts and your tone.
I would like to know if you believe your recommendations are also applicable to K-12 teachers. If so, who will teach them? When will they find the time to learn? How far behind their students (7-12, in particular) are they already?
The only people I've seen commenting about ChatGPT (any version) in education are retired K-12 teachers, educational pundits, and non-teachers. Since the program only came out in November, K-12 teachers have not had time to learn about it or learn how to cope with it, let alone adapt it to their teaching practices.