Many of Pacific’s approved AI tools have applications for learning, if used intentionally and cautiously. An upside of AI is that it is available from any location 24 hours a day, and in cases where a student is hesitant or unable to seek support from a human, equipping them with the tools to address their needs effectively with an AI can promote better outcomes; it is also worth considering what you can do to address those other barriers.
The truth is that many students are already using AI, and something we can do as educators is to help them to learn to use it as effectively as possible. A danger of the current moment is that a gap in access or training with AI could itself exacerbate inequalities, creating a new ‘digital divide.’ It is also imperative to educate them about the limitations and biases of AI.
Table of Contents
Tools
Limitations
Promoting beneficial usage
Prompting for learning applications
ToolsUniversity Information Services (UIS) has carried out a careful review of the tools they have activated and approved, and secured agreements that specify a degree of data protection and privacy (for instance, ensuring that sensitive data will not be used for training). You can read more about this process and the agreements on their AI Features FAQ page. Using tools that have not been reviewed by UIS will have none of these protections, and is prohibited.
Approved tools with particular strengths for tutoring are as follows:
- Google Gemini, particularly the ability to set up custom Gems with tailored instructions
- Box AI, which can draw upon specified materials within secure Box folders
- NotebookLM, which offers a number of multimodal study features like conceptual maps and audio summaries
For more on the features of each tool and how to enable them, see AI Resources.
LimitationsHuman educators often add significant interpretation, commentary and emphasis as they expand on written materials, conveyed by anything from sketches to speech to gesture; they also tend to pay attention to the reactions of their students, whether they look confused or confident, and which parts of the material elicit reactions. A human educator will also draw upon their past knowledge of a student, which can be a benefit or a disadvantage.
It is important to caution students about the limitations of AI tools. AIs are known to manifest bias and to give outright incorrect information. For any application you discuss, spend some time considering where inaccuracy and bias could manifest, and what steps students can take to double-check or critique answers. If you do not see a good way to mitigate potential inaccuracies, do not recommend that application. For more on the possible concerns and topics to address in class, visit General Considerations for Using AI in Teaching and Learning.
Promoting beneficial usageMany students are looking for guidance on which uses of AI are acceptable, and are hesitant to use it (or to admit to using it) when they are unsure. By providing guidance on which uses of AI are and are not accepted, you can bring AI use into an above-board space, and promote more beneficial usage. For some guidance on how to clearly communicate your AI use policies, see our AI Course Policy Guidance page.
To support students in building better habits and ways of using AI, one method is to run in-class activities that involve AI use. By doing this, you can model good prompting and critical engagement, supervise students as they experiment, and likely see some cautionary tales that will help students to be aware of potential pitfalls. It can even be beneficial in some cases to work with individual students to set up chatbots to help them with specific skills.
There is much discussion in educational spaces about the impact AI use might have on skills development, and scholarship is starting to be published on the evidence and possible approaches. What is clear is that students will still need to develop robust skills to be able to work effectively with or without AI, and the below applications offer suggestions for ways to promote skills development while working with an AI.
Prompting for Learning Applications
An AI simply asked a question about a topic is likely to quickly summarize factual information, and give an answer. However, prompting an AI to exhibit specific behaviors can make for a more effective learning experience. Students hoping to work effectively with an AI will need to consider how to craft their prompts, and take an active role in feeding back their understanding of the material. Even so, it is important to be conscious that an AI will not provide everything that a human educator can.
A strength of AI-supported study is that the same material can very rapidly be presented in different formats. Students can ask for materials to be summarized in a wide variety of different ways: as a paragraph, bullet-points, a table, metaphorical explanation, step-by-step explanation, with commentary on the culture of a discipline or underlying assumptions, and so forth.
An informal survey conducted at the University of Waterloo (reported at the session on Unsurfaced Educational Practices & Practical Tips in Graduate Supervision and Undergraduate Learning Skills Support at the 2025 Teaching with AI conference hosted by the University of Guelph) revealed that their students often use AI to help them to represent materials in ways that are more memorable or understandable to them. Examples were: ‘explain it like I’m 5 years old,’ mnemonics, and explanations in terms of personal interests (e.g. ‘relate this chemistry material to cats so I can remember it better.’). Testing out requests for materials in different formats may even help students to gain knowledge about their own learning preferences (for example, that they understand things better with lots of examples, or a table).
It is possible to ask a chatbot like Gemini or Box AI to work with specific provided materials, and to provide full referencing for all of the claims it makes. This can be beneficial for promoting reliability.
To read more about prompt structures, see Introduction to Prompting. Below are some specific ideas for tasks that an AI can fulfil.
Conceptual tutoring
Example prompt:
“Help me to understand a particular topic, or practice a skill.”
“Prioritize asking me questions to develop my understanding (“What evidence do you have to support this idea?” “How would you go about solving this?”), connect questions to key topics and define terminology. Ask me to summarize my understanding as we go, and give feedback.”
“Work from the study materials that I have attached.”
“Provide detailed referencing for every claim you make, so that I can check the original context.”
“Adopt the persona of a patient teacher.”
“Use this guide to effective tutoring to shape your approach: https://sites.highlands.edu/tutorial-center/tutor-resources/online-tutor-training/module-4/the-socratic-method/“
An AI can be prompted to ask the student questions about their current knowledge, identify gaps, provide explanations or further questioning to address them. Students can drop class documents into the Google Gemini text field, and use this as source material.
- Before we begin, pose me a question I probably won’t be able to answer yet, to focus my thinking.
- Explain the process of x like I’m 15 years old. (Then ask follow-up questions).
- I’m confused about the difference between x and y. Can you provide examples of each and explain why they aren’t the same?
- Here is my summary of x. What key points am I missing or misunderstanding?
- Give me three different examples to help me to understand this concept.
When asked a direct question, an AI will generally default to simply providing an answer. In contrast, a human study tutor will often work hard to help a student to arrive at an answer for themselves, often by asking questions, connecting the query to key topics, or modelling the application of an approach. With careful prompting, a chatbot can be cued to favor those same behaviors.
Learning experiences can be more effective when a student’s mind is ‘primed’ by having attempted to solve a problem or answer a question before instruction, such that they will then be looking for, and valuing, the understanding that will fill in what they need. For this reason, beginning a tutoring task with a tricky question or problem before instruction takes place may be effective.
A sophisticated example of a well-developed tutoring prompt is the Tutoring Prompt developed by the Wharton Generative AI labs, which outlines a multi-stage process for the AI to follow.
Practice questions
Example prompt:
“Help me to revise for a particular exam, or to practice a particular skill.”
“Pose me questions or case studies, one at a time. Give me detailed feedback on my answers. Gradually increase the difficulty as we go.”
“Work from specified study materials.”
“Provide detailed referencing so that I can check your analysis and corrections.”
“Adopt the persona of a tough examiner so that I can practice managing stress.”
“Work from these example exam questions or case studies provided by my professor.”
Chatbots can be used to generate case studies or practice questions from specific course content, give students feedback on their answers, and provide daily challenges to memorize content.
- Here is an example problem. Create a similar problem, but a little more difficult.
- Write me a set of exam questions to test my knowledge of
. Have me answer them one by one, and give feedback on my answers. - Help me to memorize this list of medications with a daily challenge: When I prompt you to, show me the list with one missing.
- Write me a case study following this example, and help me to role-play my interaction with the patient. Critique my questions and responses.
Learning is often most effective when tasks are challenging, so an interesting approach can be to prompt the chatbot to gradually increase the difficulty of the questions to keep stretching their abilities. Varying the format of the questions can also strengthen the learning.
Assignment critique
Example prompt:
“Help me to check that I have fulfilled the requirements for this assignment, and help me to notice any weaknesses in my argumentation.”
“Write a list of bullet points, one for each rubric item, detailing how well I have fulfilled each one. Then write me some paragraphs of general feedback on my clarity and communication. If there are errors in spelling or grammar, write me a list of bullet points telling me on which page and paragraph they are to be found. Finally, critique my argumentation. Suggest counter-arguments that I have not addressed. Make all of your suggestions concrete and actionable.”
“Work from this assignment description and rubric. Here is my assignment as a word document. Here are some of the relevant slides and readings from the class.”
“Provide detailed referencing so that I can check your analysis and corrections.”
“Adopt the persona of my professor.”
“Work from this example essay provided by my professor.”
A student can use an AI tool to compare an assignment against the description and rubric, and suggest areas for improvement. Specific tasks might include:
- Compare this assignment against the rubric. Write bullet-points telling me how well it fulfils each item on the rubric.
- Read a paragraph, provide feedback on clarity, and offer specific suggestions for improvement.
- Identify the main argument in a paragraph. Is it well-supported? Suggest ways to strengthen the support.
- Check a paragraph for passive voice and suggest active voice alternatives.
- Identify strengths and weaknesses of the writing style in a sample. Exemplify alternative styles.
The usefulness of feedback is often dependent on how specific and actionable it is, so it can be important to request concrete suggestions, and to push back to ask for more specificity as needed.
Idea brainstorming
Example prompt:
“Help me to come up with ideas for a 30-minute presentation for my computer science class.”
“Ask me questions about the topics in I am most interested in, and ask follow-up questions to determine my level of understanding of each. On that basis, suggest five ideas for topics for my project. For each idea, include three bullet-points identifying how each meets the assignment criteria.”
“Work from this assignment description and rubric.”
“I want to come up with something a little bit challenging and original, something different from my classmates, but I want to make sure I feel really confident with the material because they will be asking me questions at the end.”
“This is a 400-level class, so provide suggestions that are appropriate to an undergraduate at that level.”
“I am aiming for the level of this example presentation provided by my professor.”
A chatbot can work from an assignment description, ask student about their interests and existing knowledge, help draft and critique ideas.
- Help me to come up with ideas for an upcoming project, based on my knowledge and interests.
- Help me to choose between these three ideas by asking me questions about the assignment and how well these ideas fulfil it.
- Help me to choose between some possible ideas by showing me their strengths and weaknesses, and anticipate counterarguments.
- I have chosen a topic. Show me three different angles I might want to cover in my discussion.
- I am stuck writing this essay. Suggest three possible directions to take in the next paragraph.
To see a collection of prompts designed for a wide variety of tasks, visit the GenAI Chatbot Prompt Library for Students published by AI for Education.
For more guidance on how to use these tools, visit:
General Considerations for Using AI in Teaching and Learning
Course Policy Guidance
Introduction to Prompting
Assignments in the Context of AI
Labor-Saving Tips for Educators
University Information Services (UIS) has carried out a careful review of the tools they have activated and approved, and secured agreements that specify a degree of data protection and privacy (for instance, ensuring that sensitive data will not be used for training). You can read more about this process and the agreements on their AI Features FAQ page. Using tools that have not been reviewed by UIS will have none of these protections, and is prohibited.
Approved tools with particular strengths for tutoring are as follows:
- Google Gemini, particularly the ability to set up custom Gems with tailored instructions
- Box AI, which can draw upon specified materials within secure Box folders
- NotebookLM, which offers a number of multimodal study features like conceptual maps and audio summaries
For more on the features of each tool and how to enable them, see AI Resources.
LimitationsHuman educators often add significant interpretation, commentary and emphasis as they expand on written materials, conveyed by anything from sketches to speech to gesture; they also tend to pay attention to the reactions of their students, whether they look confused or confident, and which parts of the material elicit reactions. A human educator will also draw upon their past knowledge of a student, which can be a benefit or a disadvantage.
It is important to caution students about the limitations of AI tools. AIs are known to manifest bias and to give outright incorrect information. For any application you discuss, spend some time considering where inaccuracy and bias could manifest, and what steps students can take to double-check or critique answers. If you do not see a good way to mitigate potential inaccuracies, do not recommend that application. For more on the possible concerns and topics to address in class, visit General Considerations for Using AI in Teaching and Learning.
Promoting beneficial usageMany students are looking for guidance on which uses of AI are acceptable, and are hesitant to use it (or to admit to using it) when they are unsure. By providing guidance on which uses of AI are and are not accepted, you can bring AI use into an above-board space, and promote more beneficial usage. For some guidance on how to clearly communicate your AI use policies, see our AI Course Policy Guidance page.
To support students in building better habits and ways of using AI, one method is to run in-class activities that involve AI use. By doing this, you can model good prompting and critical engagement, supervise students as they experiment, and likely see some cautionary tales that will help students to be aware of potential pitfalls. It can even be beneficial in some cases to work with individual students to set up chatbots to help them with specific skills.
There is much discussion in educational spaces about the impact AI use might have on skills development, and scholarship is starting to be published on the evidence and possible approaches. What is clear is that students will still need to develop robust skills to be able to work effectively with or without AI, and the below applications offer suggestions for ways to promote skills development while working with an AI.
Prompting for Learning Applications
An AI simply asked a question about a topic is likely to quickly summarize factual information, and give an answer. However, prompting an AI to exhibit specific behaviors can make for a more effective learning experience. Students hoping to work effectively with an AI will need to consider how to craft their prompts, and take an active role in feeding back their understanding of the material. Even so, it is important to be conscious that an AI will not provide everything that a human educator can.
A strength of AI-supported study is that the same material can very rapidly be presented in different formats. Students can ask for materials to be summarized in a wide variety of different ways: as a paragraph, bullet-points, a table, metaphorical explanation, step-by-step explanation, with commentary on the culture of a discipline or underlying assumptions, and so forth.
An informal survey conducted at the University of Waterloo (reported at the session on Unsurfaced Educational Practices & Practical Tips in Graduate Supervision and Undergraduate Learning Skills Support at the 2025 Teaching with AI conference hosted by the University of Guelph) revealed that their students often use AI to help them to represent materials in ways that are more memorable or understandable to them. Examples were: ‘explain it like I’m 5 years old,’ mnemonics, and explanations in terms of personal interests (e.g. ‘relate this chemistry material to cats so I can remember it better.’). Testing out requests for materials in different formats may even help students to gain knowledge about their own learning preferences (for example, that they understand things better with lots of examples, or a table).
It is possible to ask a chatbot like Gemini or Box AI to work with specific provided materials, and to provide full referencing for all of the claims it makes. This can be beneficial for promoting reliability.
To read more about prompt structures, see Introduction to Prompting. Below are some specific ideas for tasks that an AI can fulfil.
Conceptual tutoring
Example prompt:
“Help me to understand a particular topic, or practice a skill.”
“Prioritize asking me questions to develop my understanding (“What evidence do you have to support this idea?” “How would you go about solving this?”), connect questions to key topics and define terminology. Ask me to summarize my understanding as we go, and give feedback.”
“Work from the study materials that I have attached.”
“Provide detailed referencing for every claim you make, so that I can check the original context.”
“Adopt the persona of a patient teacher.”
“Use this guide to effective tutoring to shape your approach: https://sites.highlands.edu/tutorial-center/tutor-resources/online-tutor-training/module-4/the-socratic-method/“
An AI can be prompted to ask the student questions about their current knowledge, identify gaps, provide explanations or further questioning to address them. Students can drop class documents into the Google Gemini text field, and use this as source material.
- Before we begin, pose me a question I probably won’t be able to answer yet, to focus my thinking.
- Explain the process of x like I’m 15 years old. (Then ask follow-up questions).
- I’m confused about the difference between x and y. Can you provide examples of each and explain why they aren’t the same?
- Here is my summary of x. What key points am I missing or misunderstanding?
- Give me three different examples to help me to understand this concept.
When asked a direct question, an AI will generally default to simply providing an answer. In contrast, a human study tutor will often work hard to help a student to arrive at an answer for themselves, often by asking questions, connecting the query to key topics, or modelling the application of an approach. With careful prompting, a chatbot can be cued to favor those same behaviors.
Learning experiences can be more effective when a student’s mind is ‘primed’ by having attempted to solve a problem or answer a question before instruction, such that they will then be looking for, and valuing, the understanding that will fill in what they need. For this reason, beginning a tutoring task with a tricky question or problem before instruction takes place may be effective.
A sophisticated example of a well-developed tutoring prompt is the Tutoring Prompt developed by the Wharton Generative AI labs, which outlines a multi-stage process for the AI to follow.
Practice questions
Example prompt:
“Help me to revise for a particular exam, or to practice a particular skill.”
“Pose me questions or case studies, one at a time. Give me detailed feedback on my answers. Gradually increase the difficulty as we go.”
“Work from specified study materials.”
“Provide detailed referencing so that I can check your analysis and corrections.”
“Adopt the persona of a tough examiner so that I can practice managing stress.”
“Work from these example exam questions or case studies provided by my professor.”
Chatbots can be used to generate case studies or practice questions from specific course content, give students feedback on their answers, and provide daily challenges to memorize content.
- Here is an example problem. Create a similar problem, but a little more difficult.
- Write me a set of exam questions to test my knowledge of
. Have me answer them one by one, and give feedback on my answers. - Help me to memorize this list of medications with a daily challenge: When I prompt you to, show me the list with one missing.
- Write me a case study following this example, and help me to role-play my interaction with the patient. Critique my questions and responses.
Learning is often most effective when tasks are challenging, so an interesting approach can be to prompt the chatbot to gradually increase the difficulty of the questions to keep stretching their abilities. Varying the format of the questions can also strengthen the learning.
Assignment critique
Example prompt:
“Help me to check that I have fulfilled the requirements for this assignment, and help me to notice any weaknesses in my argumentation.”
“Write a list of bullet points, one for each rubric item, detailing how well I have fulfilled each one. Then write me some paragraphs of general feedback on my clarity and communication. If there are errors in spelling or grammar, write me a list of bullet points telling me on which page and paragraph they are to be found. Finally, critique my argumentation. Suggest counter-arguments that I have not addressed. Make all of your suggestions concrete and actionable.”
“Work from this assignment description and rubric. Here is my assignment as a word document. Here are some of the relevant slides and readings from the class.”
“Provide detailed referencing so that I can check your analysis and corrections.”
“Adopt the persona of my professor.”
“Work from this example essay provided by my professor.”
A student can use an AI tool to compare an assignment against the description and rubric, and suggest areas for improvement. Specific tasks might include:
- Compare this assignment against the rubric. Write bullet-points telling me how well it fulfils each item on the rubric.
- Read a paragraph, provide feedback on clarity, and offer specific suggestions for improvement.
- Identify the main argument in a paragraph. Is it well-supported? Suggest ways to strengthen the support.
- Check a paragraph for passive voice and suggest active voice alternatives.
- Identify strengths and weaknesses of the writing style in a sample. Exemplify alternative styles.
The usefulness of feedback is often dependent on how specific and actionable it is, so it can be important to request concrete suggestions, and to push back to ask for more specificity as needed.
Idea brainstorming
Example prompt:
“Help me to come up with ideas for a 30-minute presentation for my computer science class.”
“Ask me questions about the topics in I am most interested in, and ask follow-up questions to determine my level of understanding of each. On that basis, suggest five ideas for topics for my project. For each idea, include three bullet-points identifying how each meets the assignment criteria.”
“Work from this assignment description and rubric.”
“I want to come up with something a little bit challenging and original, something different from my classmates, but I want to make sure I feel really confident with the material because they will be asking me questions at the end.”
“This is a 400-level class, so provide suggestions that are appropriate to an undergraduate at that level.”
“I am aiming for the level of this example presentation provided by my professor.”
A chatbot can work from an assignment description, ask student about their interests and existing knowledge, help draft and critique ideas.
- Help me to come up with ideas for an upcoming project, based on my knowledge and interests.
- Help me to choose between these three ideas by asking me questions about the assignment and how well these ideas fulfil it.
- Help me to choose between some possible ideas by showing me their strengths and weaknesses, and anticipate counterarguments.
- I have chosen a topic. Show me three different angles I might want to cover in my discussion.
- I am stuck writing this essay. Suggest three possible directions to take in the next paragraph.
To see a collection of prompts designed for a wide variety of tasks, visit the GenAI Chatbot Prompt Library for Students published by AI for Education.
For more guidance on how to use these tools, visit:
General Considerations for Using AI in Teaching and Learning
Course Policy Guidance
Introduction to Prompting
Assignments in the Context of AI
Labor-Saving Tips for Educators
Human educators often add significant interpretation, commentary and emphasis as they expand on written materials, conveyed by anything from sketches to speech to gesture; they also tend to pay attention to the reactions of their students, whether they look confused or confident, and which parts of the material elicit reactions. A human educator will also draw upon their past knowledge of a student, which can be a benefit or a disadvantage.
It is important to caution students about the limitations of AI tools. AIs are known to manifest bias and to give outright incorrect information. For any application you discuss, spend some time considering where inaccuracy and bias could manifest, and what steps students can take to double-check or critique answers. If you do not see a good way to mitigate potential inaccuracies, do not recommend that application. For more on the possible concerns and topics to address in class, visit General Considerations for Using AI in Teaching and Learning.
Promoting beneficial usageMany students are looking for guidance on which uses of AI are acceptable, and are hesitant to use it (or to admit to using it) when they are unsure. By providing guidance on which uses of AI are and are not accepted, you can bring AI use into an above-board space, and promote more beneficial usage. For some guidance on how to clearly communicate your AI use policies, see our AI Course Policy Guidance page.
To support students in building better habits and ways of using AI, one method is to run in-class activities that involve AI use. By doing this, you can model good prompting and critical engagement, supervise students as they experiment, and likely see some cautionary tales that will help students to be aware of potential pitfalls. It can even be beneficial in some cases to work with individual students to set up chatbots to help them with specific skills.
There is much discussion in educational spaces about the impact AI use might have on skills development, and scholarship is starting to be published on the evidence and possible approaches. What is clear is that students will still need to develop robust skills to be able to work effectively with or without AI, and the below applications offer suggestions for ways to promote skills development while working with an AI.
Prompting for Learning Applications
An AI simply asked a question about a topic is likely to quickly summarize factual information, and give an answer. However, prompting an AI to exhibit specific behaviors can make for a more effective learning experience. Students hoping to work effectively with an AI will need to consider how to craft their prompts, and take an active role in feeding back their understanding of the material. Even so, it is important to be conscious that an AI will not provide everything that a human educator can.
A strength of AI-supported study is that the same material can very rapidly be presented in different formats. Students can ask for materials to be summarized in a wide variety of different ways: as a paragraph, bullet-points, a table, metaphorical explanation, step-by-step explanation, with commentary on the culture of a discipline or underlying assumptions, and so forth.
An informal survey conducted at the University of Waterloo (reported at the session on Unsurfaced Educational Practices & Practical Tips in Graduate Supervision and Undergraduate Learning Skills Support at the 2025 Teaching with AI conference hosted by the University of Guelph) revealed that their students often use AI to help them to represent materials in ways that are more memorable or understandable to them. Examples were: ‘explain it like I’m 5 years old,’ mnemonics, and explanations in terms of personal interests (e.g. ‘relate this chemistry material to cats so I can remember it better.’). Testing out requests for materials in different formats may even help students to gain knowledge about their own learning preferences (for example, that they understand things better with lots of examples, or a table).
It is possible to ask a chatbot like Gemini or Box AI to work with specific provided materials, and to provide full referencing for all of the claims it makes. This can be beneficial for promoting reliability.
To read more about prompt structures, see Introduction to Prompting. Below are some specific ideas for tasks that an AI can fulfil.
Conceptual tutoring
Example prompt:
“Help me to understand a particular topic, or practice a skill.”
“Prioritize asking me questions to develop my understanding (“What evidence do you have to support this idea?” “How would you go about solving this?”), connect questions to key topics and define terminology. Ask me to summarize my understanding as we go, and give feedback.”
“Work from the study materials that I have attached.”
“Provide detailed referencing for every claim you make, so that I can check the original context.”
“Adopt the persona of a patient teacher.”
“Use this guide to effective tutoring to shape your approach: https://sites.highlands.edu/tutorial-center/tutor-resources/online-tutor-training/module-4/the-socratic-method/“
An AI can be prompted to ask the student questions about their current knowledge, identify gaps, provide explanations or further questioning to address them. Students can drop class documents into the Google Gemini text field, and use this as source material.
- Before we begin, pose me a question I probably won’t be able to answer yet, to focus my thinking.
- Explain the process of x like I’m 15 years old. (Then ask follow-up questions).
- I’m confused about the difference between x and y. Can you provide examples of each and explain why they aren’t the same?
- Here is my summary of x. What key points am I missing or misunderstanding?
- Give me three different examples to help me to understand this concept.
When asked a direct question, an AI will generally default to simply providing an answer. In contrast, a human study tutor will often work hard to help a student to arrive at an answer for themselves, often by asking questions, connecting the query to key topics, or modelling the application of an approach. With careful prompting, a chatbot can be cued to favor those same behaviors.
Learning experiences can be more effective when a student’s mind is ‘primed’ by having attempted to solve a problem or answer a question before instruction, such that they will then be looking for, and valuing, the understanding that will fill in what they need. For this reason, beginning a tutoring task with a tricky question or problem before instruction takes place may be effective.
A sophisticated example of a well-developed tutoring prompt is the Tutoring Prompt developed by the Wharton Generative AI labs, which outlines a multi-stage process for the AI to follow.
Practice questions
Example prompt:
“Help me to revise for a particular exam, or to practice a particular skill.”
“Pose me questions or case studies, one at a time. Give me detailed feedback on my answers. Gradually increase the difficulty as we go.”
“Work from specified study materials.”
“Provide detailed referencing so that I can check your analysis and corrections.”
“Adopt the persona of a tough examiner so that I can practice managing stress.”
“Work from these example exam questions or case studies provided by my professor.”
Chatbots can be used to generate case studies or practice questions from specific course content, give students feedback on their answers, and provide daily challenges to memorize content.
- Here is an example problem. Create a similar problem, but a little more difficult.
- Write me a set of exam questions to test my knowledge of
. Have me answer them one by one, and give feedback on my answers. - Help me to memorize this list of medications with a daily challenge: When I prompt you to, show me the list with one missing.
- Write me a case study following this example, and help me to role-play my interaction with the patient. Critique my questions and responses.
Learning is often most effective when tasks are challenging, so an interesting approach can be to prompt the chatbot to gradually increase the difficulty of the questions to keep stretching their abilities. Varying the format of the questions can also strengthen the learning.
Assignment critique
Example prompt:
“Help me to check that I have fulfilled the requirements for this assignment, and help me to notice any weaknesses in my argumentation.”
“Write a list of bullet points, one for each rubric item, detailing how well I have fulfilled each one. Then write me some paragraphs of general feedback on my clarity and communication. If there are errors in spelling or grammar, write me a list of bullet points telling me on which page and paragraph they are to be found. Finally, critique my argumentation. Suggest counter-arguments that I have not addressed. Make all of your suggestions concrete and actionable.”
“Work from this assignment description and rubric. Here is my assignment as a word document. Here are some of the relevant slides and readings from the class.”
“Provide detailed referencing so that I can check your analysis and corrections.”
“Adopt the persona of my professor.”
“Work from this example essay provided by my professor.”
A student can use an AI tool to compare an assignment against the description and rubric, and suggest areas for improvement. Specific tasks might include:
- Compare this assignment against the rubric. Write bullet-points telling me how well it fulfils each item on the rubric.
- Read a paragraph, provide feedback on clarity, and offer specific suggestions for improvement.
- Identify the main argument in a paragraph. Is it well-supported? Suggest ways to strengthen the support.
- Check a paragraph for passive voice and suggest active voice alternatives.
- Identify strengths and weaknesses of the writing style in a sample. Exemplify alternative styles.
The usefulness of feedback is often dependent on how specific and actionable it is, so it can be important to request concrete suggestions, and to push back to ask for more specificity as needed.
Idea brainstorming
Example prompt:
“Help me to come up with ideas for a 30-minute presentation for my computer science class.”
“Ask me questions about the topics in I am most interested in, and ask follow-up questions to determine my level of understanding of each. On that basis, suggest five ideas for topics for my project. For each idea, include three bullet-points identifying how each meets the assignment criteria.”
“Work from this assignment description and rubric.”
“I want to come up with something a little bit challenging and original, something different from my classmates, but I want to make sure I feel really confident with the material because they will be asking me questions at the end.”
“This is a 400-level class, so provide suggestions that are appropriate to an undergraduate at that level.”
“I am aiming for the level of this example presentation provided by my professor.”
A chatbot can work from an assignment description, ask student about their interests and existing knowledge, help draft and critique ideas.
- Help me to come up with ideas for an upcoming project, based on my knowledge and interests.
- Help me to choose between these three ideas by asking me questions about the assignment and how well these ideas fulfil it.
- Help me to choose between some possible ideas by showing me their strengths and weaknesses, and anticipate counterarguments.
- I have chosen a topic. Show me three different angles I might want to cover in my discussion.
- I am stuck writing this essay. Suggest three possible directions to take in the next paragraph.
To see a collection of prompts designed for a wide variety of tasks, visit the GenAI Chatbot Prompt Library for Students published by AI for Education.
For more guidance on how to use these tools, visit:
General Considerations for Using AI in Teaching and Learning
Course Policy Guidance
Introduction to Prompting
Assignments in the Context of AI
Labor-Saving Tips for Educators
Many students are looking for guidance on which uses of AI are acceptable, and are hesitant to use it (or to admit to using it) when they are unsure. By providing guidance on which uses of AI are and are not accepted, you can bring AI use into an above-board space, and promote more beneficial usage. For some guidance on how to clearly communicate your AI use policies, see our AI Course Policy Guidance page.
To support students in building better habits and ways of using AI, one method is to run in-class activities that involve AI use. By doing this, you can model good prompting and critical engagement, supervise students as they experiment, and likely see some cautionary tales that will help students to be aware of potential pitfalls. It can even be beneficial in some cases to work with individual students to set up chatbots to help them with specific skills.
There is much discussion in educational spaces about the impact AI use might have on skills development, and scholarship is starting to be published on the evidence and possible approaches. What is clear is that students will still need to develop robust skills to be able to work effectively with or without AI, and the below applications offer suggestions for ways to promote skills development while working with an AI.
Prompting for Learning Applications
An AI simply asked a question about a topic is likely to quickly summarize factual information, and give an answer. However, prompting an AI to exhibit specific behaviors can make for a more effective learning experience. Students hoping to work effectively with an AI will need to consider how to craft their prompts, and take an active role in feeding back their understanding of the material. Even so, it is important to be conscious that an AI will not provide everything that a human educator can.
A strength of AI-supported study is that the same material can very rapidly be presented in different formats. Students can ask for materials to be summarized in a wide variety of different ways: as a paragraph, bullet-points, a table, metaphorical explanation, step-by-step explanation, with commentary on the culture of a discipline or underlying assumptions, and so forth.
An informal survey conducted at the University of Waterloo (reported at the session on Unsurfaced Educational Practices & Practical Tips in Graduate Supervision and Undergraduate Learning Skills Support at the 2025 Teaching with AI conference hosted by the University of Guelph) revealed that their students often use AI to help them to represent materials in ways that are more memorable or understandable to them. Examples were: ‘explain it like I’m 5 years old,’ mnemonics, and explanations in terms of personal interests (e.g. ‘relate this chemistry material to cats so I can remember it better.’). Testing out requests for materials in different formats may even help students to gain knowledge about their own learning preferences (for example, that they understand things better with lots of examples, or a table).
It is possible to ask a chatbot like Gemini or Box AI to work with specific provided materials, and to provide full referencing for all of the claims it makes. This can be beneficial for promoting reliability.
To read more about prompt structures, see Introduction to Prompting. Below are some specific ideas for tasks that an AI can fulfil.
Conceptual tutoring
Example prompt:
“Help me to understand a particular topic, or practice a skill.”
“Prioritize asking me questions to develop my understanding (“What evidence do you have to support this idea?” “How would you go about solving this?”), connect questions to key topics and define terminology. Ask me to summarize my understanding as we go, and give feedback.”
“Work from the study materials that I have attached.”
“Provide detailed referencing for every claim you make, so that I can check the original context.”
“Adopt the persona of a patient teacher.”
“Use this guide to effective tutoring to shape your approach: https://sites.highlands.edu/tutorial-center/tutor-resources/online-tutor-training/module-4/the-socratic-method/“
An AI can be prompted to ask the student questions about their current knowledge, identify gaps, provide explanations or further questioning to address them. Students can drop class documents into the Google Gemini text field, and use this as source material.
- Before we begin, pose me a question I probably won’t be able to answer yet, to focus my thinking.
- Explain the process of x like I’m 15 years old. (Then ask follow-up questions).
- I’m confused about the difference between x and y. Can you provide examples of each and explain why they aren’t the same?
- Here is my summary of x. What key points am I missing or misunderstanding?
- Give me three different examples to help me to understand this concept.
When asked a direct question, an AI will generally default to simply providing an answer. In contrast, a human study tutor will often work hard to help a student to arrive at an answer for themselves, often by asking questions, connecting the query to key topics, or modelling the application of an approach. With careful prompting, a chatbot can be cued to favor those same behaviors.
Learning experiences can be more effective when a student’s mind is ‘primed’ by having attempted to solve a problem or answer a question before instruction, such that they will then be looking for, and valuing, the understanding that will fill in what they need. For this reason, beginning a tutoring task with a tricky question or problem before instruction takes place may be effective.
A sophisticated example of a well-developed tutoring prompt is the Tutoring Prompt developed by the Wharton Generative AI labs, which outlines a multi-stage process for the AI to follow.
Practice questions
Example prompt:
“Help me to revise for a particular exam, or to practice a particular skill.”
“Pose me questions or case studies, one at a time. Give me detailed feedback on my answers. Gradually increase the difficulty as we go.”
“Work from specified study materials.”
“Provide detailed referencing so that I can check your analysis and corrections.”
“Adopt the persona of a tough examiner so that I can practice managing stress.”
“Work from these example exam questions or case studies provided by my professor.”
Chatbots can be used to generate case studies or practice questions from specific course content, give students feedback on their answers, and provide daily challenges to memorize content.
- Here is an example problem. Create a similar problem, but a little more difficult.
- Write me a set of exam questions to test my knowledge of
. Have me answer them one by one, and give feedback on my answers. - Help me to memorize this list of medications with a daily challenge: When I prompt you to, show me the list with one missing.
- Write me a case study following this example, and help me to role-play my interaction with the patient. Critique my questions and responses.
Learning is often most effective when tasks are challenging, so an interesting approach can be to prompt the chatbot to gradually increase the difficulty of the questions to keep stretching their abilities. Varying the format of the questions can also strengthen the learning.
Assignment critique
Example prompt:
“Help me to check that I have fulfilled the requirements for this assignment, and help me to notice any weaknesses in my argumentation.”
“Write a list of bullet points, one for each rubric item, detailing how well I have fulfilled each one. Then write me some paragraphs of general feedback on my clarity and communication. If there are errors in spelling or grammar, write me a list of bullet points telling me on which page and paragraph they are to be found. Finally, critique my argumentation. Suggest counter-arguments that I have not addressed. Make all of your suggestions concrete and actionable.”
“Work from this assignment description and rubric. Here is my assignment as a word document. Here are some of the relevant slides and readings from the class.”
“Provide detailed referencing so that I can check your analysis and corrections.”
“Adopt the persona of my professor.”
“Work from this example essay provided by my professor.”
A student can use an AI tool to compare an assignment against the description and rubric, and suggest areas for improvement. Specific tasks might include:
- Compare this assignment against the rubric. Write bullet-points telling me how well it fulfils each item on the rubric.
- Read a paragraph, provide feedback on clarity, and offer specific suggestions for improvement.
- Identify the main argument in a paragraph. Is it well-supported? Suggest ways to strengthen the support.
- Check a paragraph for passive voice and suggest active voice alternatives.
- Identify strengths and weaknesses of the writing style in a sample. Exemplify alternative styles.
The usefulness of feedback is often dependent on how specific and actionable it is, so it can be important to request concrete suggestions, and to push back to ask for more specificity as needed.
Idea brainstorming
Example prompt:
“Help me to come up with ideas for a 30-minute presentation for my computer science class.”
“Ask me questions about the topics in
“Work from this assignment description and rubric.”
“I want to come up with something a little bit challenging and original, something different from my classmates, but I want to make sure I feel really confident with the material because they will be asking me questions at the end.”
“This is a 400-level class, so provide suggestions that are appropriate to an undergraduate at that level.”
“I am aiming for the level of this example presentation provided by my professor.”
A chatbot can work from an assignment description, ask student about their interests and existing knowledge, help draft and critique ideas.
- Help me to come up with ideas for an upcoming project, based on my knowledge and interests.
- Help me to choose between these three ideas by asking me questions about the assignment and how well these ideas fulfil it.
- Help me to choose between some possible ideas by showing me their strengths and weaknesses, and anticipate counterarguments.
- I have chosen a topic. Show me three different angles I might want to cover in my discussion.
- I am stuck writing this essay. Suggest three possible directions to take in the next paragraph.
To see a collection of prompts designed for a wide variety of tasks, visit the GenAI Chatbot Prompt Library for Students published by AI for Education.
For more guidance on how to use these tools, visit:
General Considerations for Using AI in Teaching and Learning
Course Policy Guidance
Introduction to Prompting
Assignments in the Context of AI
Labor-Saving Tips for Educators