Make the most of prompt writing | University at Albany

Making the most of prompt writing for learning and teaching

When you decide that you want students to use artificial intelligence in productive, pedagogically sound ways, it is important to fully explore how you can ensure that they make the most of that technology for their learning. One key part of that exploration means recognizing that when they use AI, what is really occurring is that AI is responding to the prompts that our students design. If we don’t focus on prompt design itself, we miss a key piece of learning that can potentially benefit our students.

In this short guide, we make some suggestions about exploring prompt writing with your students so that they can use this aspect of interacting with AI to learn and grow. Then we provide examples of the kinds of prompts that work best when students use AI to support their learning. Finally, we consider how prompt writing can benefit our students cognitively and metacognitively.

Exploring prompt writing with your students

Students will need guidance to develop or use prompts to get the most out of their interactions with AI. Here are a few suggestions about how to support students as they learn about prompt design.

Be clear about why and how you want students to use AI

Before you engage with exploring prompt writing with your students, you will want to do some initial thinking about how and why you want students to use AI in your course. Assuming you’ve done this careful thinking and determined that using AI will help students achieve the goals you’ve set for them, you will also want to be sure you’ve communicated how and why they will be using AI. These conversations are important: research suggests that our students are eager to understand more about the ethical uses of AI and to understand the role it will play in their learning at the university (Digital Education Council, 2024). One important aspect of these conversations is also what you don’t want them to use AI for. Reaching an understanding together of when, why, and how to use AI effectively sets the stage for meaningful learning with AI.

Help students understand what makes prompts more and less effective

Spend some time with students helping them understand how AI works when it is prompted. This can be done in many ways. You can have students critique and explore the generated texts or images that AI generates when given basic prompts like, “Draw a picture of a cat and a girl” or “What is good picnic food?”. Have students decide what biases are reflected in the results. Have them also consider how and why the results are limited or constrained. These conversations will set the stage for a discussion about how AI really works. Helping students understand that AI is not thinking but is rather using calculated probable relations between words or other semiotic bits of information it has been trained on allows students to see that without much greater detail in a prompt, the results are a kind of average of averages—often quite bland as well as potentially biased and limited.

You could next have them consider what will make prompts more productive for their learning in your course. One powerful way to do this is to share a less and more detailed prompt with students. For example, an instructor teaching an introduction to linguistics course might share these two prompts with her students as well as the results to help them see how they can write detailed prompts to use AI as a thinking partner as they do their reading for the course:

Less detailed prompt: “What is the Sapir-Whorf hypothesis?”

More detailed prompt: “I’m an undergraduate taking an introduction to linguistics course. I don’t really understand the Sapir-Whorf hypothesis. Please give me an explanation I can understand. I also am not sure why this hypothesis is important. What is at stake? Please also give me two examples of it. I am from a Dominican family and grew up speaking Spanish at home as well as English. The examples we are given in our book are only about tribes in South America. Can you give me some examples of the hypothesis that are about Spanish language?”

Have students reflect on their work with AI

Depending on how you choose to have students use AI in your course, you may have them writing and developing prompts with you or with peers. You may also be providing students with prompts that you have them analyze or edit with your supervision. Regardless of how prompts are generated, students need to work with the results of what AI produces so that they continue to reflect on the exchange they’ve co-created with it. This means that after you’ve explored prompt writing with your students, you should have them make decisions about how they will share their prompts with you and how they will reflect on the work that they have done with AI. Students, with your support, can create a template for these reflections that can include the prompts they used and a short paragraph describing what they learned, where they still need support, and what next steps they will take.

It is also helpful to have students critically reflect on their developing understanding of AI as a partner in their learning. These reflective addenda to assignments where students used AI should be an opportunity for them and for you to consider the benefits and the drawbacks of the mindful use of artificial intelligence by humans. Research suggests that without this kind of metacognitive reflection, our students may easily become less critical thinkers and simply lean into AI in ways that deplete the very mental capacities we hope to develop as their instructors (Lee, Sarkar, Tankelevitch, Drosos, Rintel, Banks, & Wilson, 2025).

The next section of this resource has examples of prompts that students might use to interact with AI in different ways: as a quizzing partner, as a thinking partner, and as a fallible partner. Each example also describes reflective work that students could do to ensure that they are thinking critically about their use of AI and the results it produces.

Examples of prompts that can support student learning

When students prompt AI to support them in their learning, they are likely doing one of three things: asking AI to help move their thinking forward, asking AI to quiz them in some way about their current state of understanding, or exploring AI itself to discover what it can or can’t do in relation to some disciplinary standards. Let’s take a look at examples of three prompts that would help students do this kind of work with AI. Note that these examples come from a diverse range of disciplines: a STEM discipline, a social science discipline, and a humanities discipline. It should be apparent that these kinds of prompts may be productive for students across the disciplines.

Example 1: Using AI as a quizzing partner

One way that you may want students to use AI is as a quizzing partner. Students will need support in learning how to prompt AI and how to make the most of the responses they get. Let’s dig into this with an example of a prompt to help students in a sociology course quiz themselves on their understanding.

“I am an upper-level undergraduate sociology student, and I am trying to improve my research analysis skills. I read the uploaded article. Please ask me questions about the methods, statistics, findings, and limitations in this article. Ask me one question at a time and make each question short-answer or multiple choice. After I answer a question, give me feedback on my answer. Then ask me if I am ready for my next question. When I get a question wrong, make the next question a little easier. When I get a question right, make the next question a little harder so that you are testing more than just whether I have memorized information from the article. Keep asking me questions one by one until I ask you to stop. When I ask you to stop, give me a summary of which parts of the article I understand best and which parts I have confusion about.”

After this experience, students submit a short reflective essay to their instructor where they articulate one area in reading research where they have made some progress and one area where they are still struggling. They articulate a question that they will bring with them to class to help them in the area where they are still feeling challenged. They also write a few sentences about their experiences using AI, if they feel that this use was productive for them, and how they might continue to use AI as a quizzing partner in the future.

Example 2: Using AI as a thinking partner

Another way that you may want students to use AI is as a thinking partner. Students will need support in learning how to create prompts that ensure AI does not do their work for them, but rather gives them suggestions and feedback on what they have done or gets them started in some way. They will also need support in learning how to make the most of the responses they get. Let’s dig into this with an example of a prompt to help students in a physics course get feedback to improve their thinking.

“I am an undergraduate student in an introduction to physics course. I would like help thinking about and improving my lab report. I don’t want you to rewrite my lab report, but I want some feedback about two areas. Please focus on my results section and my conclusions section. I want to provide good evidence and details from my experiment in these sections to support the main points I am making. Please point out areas where I am doing this well and areas where I need more detail from my experiment.”

Students share their AI prompts at the end of their lab report so that the instructor can see how they used AI. Then students write a sentence or two summarizing the feedback they got, what edits they made based on this feedback, and what they learned that they can use on their own to help them write stronger reports in the future. They might also be asked to share ideas about this experience with AI and how they could use AI as a thinking partner in the future in this course or in the real world of research in STEM.

Example 3: Using AI as a fallible partner

Finally, you may want students to work with AI as a fallible partner, a technology that can produce text or images, but ones that need to be critiqued and evaluated. This might be the case if AI is used in your discipline and you want students to become adept at exploring and critiquing what AI generates. In these cases, your students will need support in learning how to prompt AI in particular ways to explore the limits of what it can produce with and without deeper disciplinary guidance. This is important because students will need to recognize that the less expert prompting they give AI, the less precise and potentially misleading or less useful the generated text or images will be. Students will also need support in considering how to critique or evaluate the responses they get. Let’s dig into this with an example of a prompt to help students in a humanities course critically explore AI.

“I am a student in an undergraduate sculpture class. We are learning about installation art. Please design a museum space for a set of large sculptures made by an installation artist who creates objects that range in size from two to twenty feet in diameter. The objects are constructed from discarded toys and plushies. I have uploaded an image of one of the objects which is ten by 5 feet. Design the space as well as the arrangement of the pieces.”

Students then evaluate the installation space generated by AI using aesthetic, cultural, and critical theories they are studying in their course. Students formalize their evaluation in writing, describing their new understandings of both course concepts and the limitations and potential of AI for artists and for those who design spaces for them.

How prompt writing can benefit students cognitively and metacognitively

As you reflect on the three examples above, you will notice that the prompts have been designed very carefully, with a great deal of detail. Productive prompts for learning are designed with two things in mind: the goals that students have for their learning and their emerging knowledge of the discipline and the course itself. Requiring students to draw on their goals and knowledge as they create a detailed draft for a prompt means that they are doing a lot of meaningful thinking.

Prompt writing and cognitive development

Let’s consider how students’ goals for their learning get integrated into prompts. In the first two examples, using AI as a quizzing partner and as a thinking partner, the prompts are written to ensure that AI is not thinking for the student or doing the work for the student. The prompts are written to ensure that the student is using AI to push their thinking and understanding forward. A great deal of detail is in the prompt to ensure that the student gets just the right amount of support. In the third example, using AI as a fallible partner, the prompt is written purposefully with less detail to ensure that AI generates something less predictable for the student to critique.

Another feature of these prompts is a focus on the processes, concepts, and skills that students are learning in the course. In all three prompts, there is detail drawn from what and how students are learning as well as specific concepts. When students write productive prompts to learn with AI, they must focus on what they are learning and articulate the questions and support they need to move forward with that learning. And this also helps us consider that good prompt writing necessarily requires students to reflect on the learning process itself.

Prompt writing and metacognitive development

Metacognitive work such as monitoring their knowledge, articulating where they are struggling, where they need help, or the next steps they need to take is highly valuable work for our students. If we want students to use AI to support their learning, we can help them use this metacognitive work to deepen their learning in our course as well as form more self-regulatory skills that will last beyond our courses.

As you consider these three examples, notice that thoughtful work with prompts entails a deep and focused use of course principles, disciplinary thinking, and reflection. In all three examples, we have described work students should do after AI has responded to the prompts. This kind of reflective work helps students take control of and learn from the experience they have had with AI. In the third example, students work in a very focused way with what AI generates, critiquing what is generated with key course concepts. When we have students engage in this kind of work, we are ensuring not only that they learn to make the most of their interactions with AI, but also that they use those interactions to develop into more mature thinkers.

Resources

Alby, C. (2023). ChatGPT: A Must-See Before the Semester Begins, Faculty Focus. https://www.facultyfocus.com/articles/teaching-with-technology-articles/chatgpt-a-must-see-before-the-semester-begins/
Bowen, J. A., & Watson, C. E. (2024). Teaching with AI: A Practical Guide to a New Era of Human Learning. Johns Hopkins University Press.
Lee, H., Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (April 26-May 1 2025). The impact of generative AI on critical thinking: Self-reported reductions in cognitive effort and confidence effects from a survey of knowledge workers [Conference presentation]. CHI Conference on Human factors in Computing Systems, Yokohama, Japan.