The semester is almost over. Spend one quiet hour reviewing your AI teaching practices with these questions, and you’ll walk into January with a plan instead of a pile of guesses.
Answer them honestly and write your answers down. By the time you finish, you’ll know which AI habits earned a place in your classroom, which ones need fixing, and what your students should hear from you on the first day back.
1. Did My Students Know When AI Was Allowed?
Think back to the AI question students asked you most this semester. Odds are it was some version of “Can I use it on this?” When students keep asking what the rule is, you haven’t stated it clearly, and each student has to guess what you expect.
Unclear rules hurt honest students the most. A careful student avoids AI completely and spends 3 hours on work a classmate finished in 30 minutes. Another student uses it on everything, assuming your silence meant permission. When a problem finally shows up, you’re stuck enforcing a rule nobody ever saw in writing.
Write the AI rule into each assignment
A policy in the syllabus helps, but it can’t cover every task. A brainstorming exercise and a final essay need different rules. A short AI statement on each major assignment fixes this by answering:
What’s allowed. For example, “You may use AI to generate topic ideas and to check grammar.”
What has to be disclosed. Ask students to name the tool they used and paste the prompt they wrote.
What they still have to do on their own. Name the thinking you want to see, such as choosing the evidence or explaining why a solution works.
The last item matters most to students. When they know what they will have to explain without help, they can better judge when AI is helping them learn and when it is doing the work for them.
Use a simple scale students can remember
Some teachers label every assignment with 1/4 levels: no AI, AI for planning only, AI with disclosure, or open use. A shared scale lets students check the label instead of raising a hand. If your whole department agrees on the same scale, students won’t have to learn a new set of rules in every class they take.
Fix this semester’s assignments first
Pull up three assignments from this fall where AI caused confusion. For each one, write the AI statement you wish you had included. Those three statements become your template for January, and you’ll spend the first week of the new semester teaching instead of answering the same question 20 times.
2. Did My Assignments Still Show Me How Students Think?
Remember the stack of essays that came in clean, organized, and nearly identical in tone? Each one earned a decent grade, and you still couldn’t say what any single student had learned. That’s the clearest sign an assignment has stopped doing its job. When AI can produce the finished work, the work no longer shows you what the student who turned it in understands.
Banning AI won’t solve this, because you can’t reliably detect it, and detection tools have flagged plenty of honest students. A better approach is to grade more of the thinking that happens before and after the final draft.
Find the assignments AI can finish alone
Go back through this semester’s major assignments and ask one question about each: could a student with a good prompt complete this without understanding the material? Summaries, generic five-paragraph essays, and short-answer worksheets usually fail that test. Assignments tied to class discussions, local examples, or a student’s own experience usually pass, because AI doesn’t know what happened in your room on a Tuesday in October.
Redesign them so reasoning shows up
You don’t have to rebuild everything. Small changes to an existing assignment can bring student thinking back into view:
Collect the drafts. Ask for an outline, a first draft, and a final version, and grade the changes between them.
Add a short conversation. A two-minute check-in where students explain one choice they made tells you more than a page of polished writing.
Ask for a reflection. Have students write a paragraph about what they changed, what they rejected, and why.
Make them defend a revision. Give students an AI-generated answer with flaws in it and ask them to find the problems and fix them.
Each of these changes lets you see how the student reasoned, and that is what you need to grade.
Start with one assignment
Pick the assignment that frustrated you most this fall and add one of these changes before January. Run it once, see what it shows you about your students, and then decide which assignment to adjust next.
3. Did AI Save Me Time on the Work That Matters Most?
Look back at the ways you used AI for planning and preparation this fall, and put each one into a list:
Tasks where it helped. These are jobs like drafting practice questions, rewording directions for different reading levels, or generating examples you could pick from.
Tasks where it cost you. These are jobs where you spent more time fixing the output than you would have spent doing the work yourself.
The second list usually has one thing in common. You gave the tool a vague request, and it filled in the blanks with guesses.
Write a better lesson-planning brief
A clear brief keeps your judgment in charge and cuts down on repairs. Before asking AI for help with a lesson, write out:
The objective. State exactly what students should know or be able to do by the end.
The audience. Describe the grade level, reading level, and anything you know about this particular group.
The constraints. Include class length, available materials, and the standard you’re covering.
The review criteria. Decide ahead of time how you’ll judge whether the draft is usable. This gives you a quick way to accept, revise, or throw out what comes back.
Keep what worked
Save the briefs that produced good results. Reusing a strong brief in January will take minutes, and you’ll spend your planning time on the decisions only you can make about your students.
4. Did I Check AI’s Work as Carefully as I Check My Students’ Work?
You would never hand back a student essay without reading it. Yet plenty of teachers this fall printed an AI-generated worksheet, summary, or answer key after a quick skim, and a few of them found the mistake only when a student pointed it out in class. That moment stings, and it costs you some credibility with the students you’re asking to use AI responsibly.
AI tools write with confidence whether they’re right or wrong. They can invent a quote, misstate a date, cite an article that doesn’t exist, or explain a math step in a way that skips the reasoning. None of that looks wrong at a glance, which is exactly why a quick skim isn’t enough.
Build a five-minute review habit
Before any AI-generated material reaches your students, run it through 4 checks:
Sources. Confirm every quote, statistic, and citation against the original. If you can’t find the source, cut the claim.
Accuracy. Work through answer keys and examples yourself, especially in math and science, where one wrong step can confuse a whole class.
Accessibility. Check reading level, font, layout, and directions so every student, including those with accommodations, can actually use the material.
Fit. Ask whether the material teaches the objective you set, or whether it drifted toward something easier to generate.
Most materials will pass in a few minutes. The ones that don’t would have cost you far more time after they reached your students.
Show students how you check
Your review habit can become a lesson. Put an AI-generated paragraph on the board, admit that you needed to fix it, and walk students through how you found the problems. Then give them a flawed AI answer and ask them to check it the same way.
Students learn verification faster when they watch you do it than when they hear a rule about it. By the end of next semester, you’ll have a class that questions what AI tells them, and you’ll have earned their trust by holding your own materials to the same standard you hold their work.
5. Did I Share What Worked With My Colleagues?
Think about the best AI discovery you made this semester. Maybe it was a prompt that turned a dense reading into 3 versions for different reading levels, or an assignment statement that finally stopped students from asking the same question every week. Now think about how many people in your building know about it. For most teachers and professors, the honest answer is one or two, and that’s usually because of a hallway conversation that ended before anyone wrote anything down.
Meanwhile, your colleague next door spent October solving the same problem from scratch. Across a department, that adds up to dozens of hours spent rediscovering things someone else already figured out.
Start a small shared library
You don’t need a committee or a new platform. A shared folder or a single document is enough to get started. Label everything with the class level and subject so colleagues can find what fits their classes.
Share the failures too
The prompts that went wrong are just as useful as the ones that worked. If an AI-generated quiz had three wrong answers, a colleague who reads that note will check their own quizzes more carefully. Honest notes about what failed save people from repeating the same mistakes.
Each of these five questions comes back to the same idea. AI belongs in your classroom on your terms, with clear rules your students can read, assignments that still show you how they think, planning help that saves real time, materials you’ve checked before anyone else sees them, and ideas your colleagues can pick up and use. None of that requires you to become a technology expert; just an hour of honest review now, while this semester is fresh and the next one hasn’t started.
Some teachers will finish that hour with a clear plan. Others will finish it with a longer list of questions than they started with, especially if their school hasn’t settled on a policy, their department is split on how much AI to allow, or they simply ran out of energy somewhere around finals week. That’s a normal place to be in December, and it’s also a hard place to work through alone.
That’s where we can help. We work with educators on the practical side of AI in the classroom: writing assignment-level AI expectations, redesigning assignments so student thinking stays visible, building lesson-planning briefs that save time without lowering quality, and setting up review habits teachers and students can both use. We also help schools turn scattered experiments by individual teachers into shared practices a whole department can follow.
If you’d like to talk through your answers to these five questions, or figure out where your school should start before January, book a call with John. Bring the assignment that frustrated you most this fall, and we’ll start there.