Skip to content

Educational Trends, News & Teacher Insights

Updated November 16, 2026

Q&A

How Classrooms Can Set Clear Expectations for Artificial Intelligence Use

Artificial intelligence can support learning without replacing student judgment. Experts in education and AI share practical ways to set clear expectations, from defending tool choices to submitting outlines and process notes. These strategies help teachers assess real understanding through timed drills, decision memos, error reflections, and student explanations.

ENEducation News ·

§ How this was made

The perspectives in this article were written by the named contributors and submitted through Connectively. Connectively's automated system selected them and used AI to write the headline, introduction, section headings, summary sections, and illustration, and to make minor edits for clarity. Articles are assembled and published automatically. Report errors to corrections@connectively.us. How we make our content

Illustration generated with AI

What the experts said

7 signed answers

KKECJCNJDRVSDM

7 educators answered this one

  1. Annotate and Defend Tool ChoicesKuldeep Kundal
  2. Submit Outlines and Explain RevisionsEileen Chin
  3. Use Timed Drills to Prove MasteryJim Cook
  4. Add Process Notes to Build JudgmentNafis Jaleel
  5. Pair Drafts With Decision MemosDr. Richard Hudnett Ed.D., MBA
  6. Write Error Reflections Before FeedbackVaibhav Sharma
  7. Commit First Then Seek ExplanationsDr Muhammad Salar Raza

Annotate and Defend Tool Choices

KKKuldeep KundalFounder & CEO, CISIN

Protecting learning in the age of generative AI requires shifting our focus from the final artifact to the architectural process of decision-making. We must set the expectation that AI functions as a research assistant or sounding board, never the primary author. In my work mentoring entrepreneurs, I've observed that the individuals who succeed are not those who copy-paste AI-generated market analysis, but those who can defend the specific logic behind the selections the machine made. Student value now lies in synthesis and critical judgment rather than the mere retrieval of information.

To make this thinking visible, I have implemented a Critical Iteration Log. Instead of submitting a standard essay, students provide their final work alongside a transcript of their interactions with the AI. In this log, the student must annotate the machine's contributions, specifically noting where they disagreed with the tool, where they corrected a hallucination, and why they chose one specific phrasing over another. This forces the student into the role of lead architect. By requiring this metadata, an instructor can evaluate the student's logic in real-time.

This approach moves the assessment away from a polished, potentially automated result and toward the human-led critique of that result. It acknowledges that AI is a permanent fixture in modern careers while ensuring that the core skills of skepticism, logic, and structural planning remain sharp. It turns the assignment into a lesson in design thinking, where the student manages the technology rather than being managed by it. If a student cannot explain why a specific paragraph belongs in their work, they haven't completed the assignment, regardless of how well-written the prose appears.

Submit Outlines and Explain Revisions

ECEileen ChinEnglish Education Content Specialist, Write Edge

I'm Eileen Chin, an English Education Content Specialist with an English and Writing Specialist Centre in Singapore. I work in English education and curriculum development, supporting learners through structured language and writing programmes.

I believe AI should be treated as a learning tool, not a replacement for a student's own thinking. I set clear expectations that students must first understand the topic, develop their ideas and attempt the work independently before using AI for support. They should also be able to explain their answers in their own words. AI can help with brainstorming or checking grammar, but students should not submit generated content as their own work. In English and writing lessons, I believe it is especially important to develop independent thinking, vocabulary and the ability to express ideas clearly.

One assignment change that works well is asking students to submit their original outline and first draft before making any improvements. They can then use AI to review grammar or suggest ways to improve clarity, but they must explain which suggestions they accepted or rejected and why. This makes their thinking visible and encourages them to evaluate information rather than simply copy it. This approach also reflects the importance of guided practice, individual feedback and independent writing in English education.

Use Timed Drills to Prove Mastery

As a music educator, I have realized that the more theoretical something is, the more students might rely on AI instead of their own thinking. If I want to know what the A Mixolydian scale is, AI can spit that right out. But to make a great song in A Mixolydian is a different story.

So, I expect that AI can help with research, clarification, or providing potential paths forward, but the student needs to be able to both explain the reasoning in their own words and show tangible skills of competency in an AI-free environment. If they cannot do that, even if whatever was submitted was polished, then I am keen to question if much actual learning happened.

One assignment change that helped students to show their own thinking was to drill on a subject with a time constraint. Instead of focusing on the finished product, let's get our hands dirty and see the work being done. In the song example from earlier, let's put some twists in there: write a counter-melody to the original song, change up the chords to feel more mysterious, reset your compressor settings and walk me through how you made your decisions with which knobs to turn. AI is great for tossing out sometimes-correct answers, but learning requires locking both information and behaviors into the brain.

Add Process Notes to Build Judgment

NJNafis JaleelAI Strategist & Prompt Engineer, Nafis Jaleel

The expectation I teach is simple. AI can help you explore, but it cannot do your thinking for you. Students should be allowed to use it openly, with one clear rule. Show your work. Name what you asked, what the tool gave back, what you checked, and what you changed in your own words. That honesty protects learning, because the teacher can see the thinking, not just the final answer. One assignment change makes that real. Require a short process note with every submission, listing the prompt used, one claim the student verified from another source, and one choice the student made that the tool did not make for them. The technology is already part of their lives. Our job is to train judgment alongside it, so they learn to question the output instead of copying it.

Nafis Jaleel, AI Strategist and Prompt Engineer, Lake Worth Beach, Florida (nafisjaleel.com)

Pair Drafts With Decision Memos

I tell my students upfront that AI is a tool in the room, not the author of their work. In my online business and marketing courses, students may use AI to brainstorm or to draft, but the grade is on their own thinking — what they kept, what they changed, and why. The assignment change that made the biggest difference was splitting one written analysis into two parts: the AI-assisted draft, and a short memo where the student explains what they accepted from the AI, what they rejected, and what evidence or reasoning supports their final recommendation, applied to a real local business. That memo is where I can actually see their thinking. It also makes the work harder to fake and more useful to them, because they have to defend a decision instead of just submitting polished text. My expectation is simple and I repeat it every term: use the tool, show your work, and own the conclusion.

Write Error Reflections Before Feedback

VSVaibhav SharmaFounder and Certified Test Prep Expert, PrepGenix

I teach the Digital SAT and ACT one-on-one, and the AI question comes up often. My rule for students is that AI can explain, but it can't answer for you. You try the question on your own first. Then you can ask AI to explain an idea you didn't understand. There's no AI in the room on test day, so an answer it hands you during practice is a point you haven't learned.

I tell students this openly instead of banning AI. They use it for everything else, and a ban just means they use it quietly.

The change that helped most: for every question a student misses, they write one line before they see the explanation. Why did I pick that answer, and what will I check next time? AI can't write that line for them, because it doesn't know what they were thinking when they chose B. Those lines also show me the problem quickly. A student who writes "I didn't read the last sentence" needs different help from one who writes "I didn't know the formula."

Commit First Then Seek Explanations

I build exam prep for doctors, and the rule I set is "AI can explain, but it can't answer for you." Learners attempt a question first and commit to an answer, and only then use AI to understand why they were right or wrong. The change that worked best was asking students to write one line explaining their reasoning before they see the explanation. It makes the thinking visible, and the gap between what they wrote and the correct reasoning is where the learning happens.