Prompt the Human Before You Prompt the Machine

As artificial intelligence (AI) continues to evolve, our instructional practices need to be equally as agile.

In many ways, AI feels different from other technology tools we've adopted over the years. It can generate ideas, explain concepts, summarize information, provide feedback, create images, revise writing, solve problems, and respond to questions in ways that feel conversational. And, like most technology adoptions, it's easy to focus on the tool first. 

• What can Brisk do?
• Which tool is better for this task—Gemini or NotebookLM?
• What is the best prompt framework to use?

These are important questions. But, they're not the most important questions.

One of the most helpful ideas guiding the work of the AI Task Force and District Technology Committee in Barrington 220 comes from Dr. Sabba Quidwai, who provided the keynote and several sessions at our February 13, 2026, Institute Day. Her session, Design Thinking for Schools: Prompting and Partnering with AI, helped frame AI not as something that replaces human thinking, but as something that can extend it.

Sabba often uses the phrase, "Prompt the human before you prompt the machine." 

That idea sounds simple, but it changes everything. 

It reminds us that the thinking we do before using AI may be more important than the prompt we type. 

Sabba’s work around human-centered AI and the SPARK framework emphasizes that AI is most powerful when it begins with context, purpose, human judgment, and a clear understanding of the problem we are trying to solve.

That same principle is now reflected in the Barrington 220 AI Skills framework. 

Our AI Skills document names this clearly: before using AI tools, students should engage in their own thinking through brainstorming, questioning, or hypothesizing. AI should be used to extend human thinking, not replace the initial spark of inquiry.

That is the shift.

Instead of starting with:

“What prompt should I give the AI?”

We might start with:

“What question should I ask the student?”

The goal is not simply for students to become better at prompting a machine. The larger goal is for students to become better thinkers, creators, collaborators, and learners. A strong AI prompt can produce a useful response. A strong human prompt can produce something even more valuable: reflection, revision, curiosity, ownership, and deeper learning.

When students use AI well, they are not simply asking it to complete work for them, they are using it to support the thinking they are already doing. They might use AI to brainstorm possibilities, test an idea, receive feedback, consider another perspective, or reflect on how their work has changed.

For example, a student could ask AI to:

“Write my paragraph about the theme of courage.”

That prompt may produce a paragraph. It may even produce a pretty good paragraph. But the thinking has moved away from the student.

A student might instead ask:

“Here is my paragraph about the theme of courage. Can you ask me three questions that would help me make my thinking clearer?”

That prompt does something different. It keeps the student in the learning process. It asks AI to serve as a thinking partner rather than an answer machine. The difference is important. This is also why the Barrington 220 AI Skills framework describes AI as a teammate, not the pilot. AI can generate ideas, offer feedback, and analyze information, but humans bring perspective, values, creativity, and judgment. Human judgment guides the work. That difference matters.

This is not about limiting AI use. It is not about making learning harder just because a new tool exists. It is also not about pretending students will never use AI to help them complete work. Instead, this is about designing learning experiences where AI supports student thinking rather than replacing it. This idea may be one of the most important instructional design challenges in front of us.

One helpful way to think about this shift is:

Less: “AI, do this for me.”

More: “AI, help me think through this.”

In Barrington 220, we are beginning to describe this through five core AI skills:

Ask. Check & Choose. Correct. Create. Connect.

These skills give students—and teachers—a repeatable workflow for AI-supported learning. Students learn to define what they need before typing, decide what to trust or reject, revise while maintaining ownership, use AI to generate options, and connect ideas across subjects and contexts. That workflow keeps the human learner at the center.

In a classroom, this framework might look like a student using AI: 

• before writing to generate questions that help them brainstorm stronger ideas.
• during writing to receive feedback on whether their evidence supports their claim. 
• after completing a project to reflect on what changed from the first version to the final version.

In each of those examples, AI is not removing the productive struggle from learning. It is helping make that struggle more visible, supported, and useful. That distinction is important because authentic learning is not just about the final product students create. It is also about the process students use to get there. When students make decisions, revise their thinking, explain their choices, and reflect on their growth, they are doing the work of learning.

AI can support that process, but only if we design for it.

A Barrington 220 example of this can be seen in a Grade 5 persuasive writing lesson at Roslyn Road Elementary. Students first wrote persuasive essays about disaster preparedness through a teacher-led process. Then, with close teacher guidance, they used AI tools such as Snorkl and Brisk to receive feedback on their ideas, organization, and writing conventions. The technology supported the feedback loop, but the students and teachers drove the learning. That is the kind of AI use we want to continue exploring.

The goal was not for AI to write for students, the goal was for students to use feedback, make decisions, revise their work, and maintain ownership of their voice:

• In a science class, students might ask AI to challenge their explanation of a lab result by identifying possible gaps or alternative explanations. 
• In a social studies class, students might ask AI to help them examine an issue from multiple perspectives before developing their own claim. 
• In an elementary classroom, students might use AI with teacher guidance to generate questions they still have after reading or exploring a topic.

The common thread is not the tool; the common thread is the thinking.

As we continue to explore student-facing AI tools, we can look for evidence that students are still active participants in the learning process. We might look for students:

• explaining why they used AI.
• revising based on feedback.
• asking for questions instead of answers.
• comparing AI feedback to success criteria.
• reflecting on what they learned through the process. 

These are the kinds of classroom look-fors that help us move beyond whether AI was used—and toward whether learning was strengthened.

A simple test for students could be:

“Will this help me think, create, revise, or reflect?”

If the answer is yes, AI may be supporting learning.

If the answer is no, and AI is only helping skip the thinking, then we may need to redesign the task, the prompt, or the expectations.

Teachers might also try a few student-centered AI prompts that keep the learner in control:

• “I am working on _____. Do not give me the answer. Ask me one question at a time to help me think through the problem.”
• “Here is my work. Give me feedback on _____. Tell me one strength, one area to improve, and one question I should consider.”
• “Help me reflect on my learning. Ask me questions about what I tried, what changed, what I still wonder, and what I would do differently next time.”

These prompts are simple, but they reinforce an important idea: the student is still the learner, the thinker, and the decision-maker. As AI becomes more common in our classrooms and in the world beyond school, students will need more than technical skills. They will need judgment. They will need curiosity. They will need the ability to ask better questions, evaluate responses, revise their work, and explain their thinking. That is why our work cannot stop at teaching students how to prompt AI.

We also need to help students prompt themselves.

• What am I trying to learn?
• What do I already understand?
• Where am I stuck?
• What feedback would help me improve?
• What decision do I need to make next?
• How do I know this work reflects my thinking?

Those questions are not new. They have always been part of strong teaching and learning. AI simply gives us a new reason to make them more visible. If AI can complete parts of a task quickly, then we have an opportunity to ask an even better instructional question:

What is the learning we most want students to do?

Dr. Sabba Quidwai’s visit reminded us that design thinking and AI belong together because both begin with human purpose. Our Barrington 220 AI Skills framework gives us a practical way to bring that idea into classrooms through Ask, Check & Choose, Correct, Create, and Connect. 

The future of AI in classrooms is not just about better prompts for machines, it is about better prompts for learners. And that feels like the work that matters most.