Skip to content

Designing AI-Resistant Questions That Inspire Real Thinking

Like teachers all over the world, I’ve been trying to figure out when we should and shouldn’t let students use AI, and how we can still ensure authenticity of work when we’re not watching them do it.

I used to think it was important to have students repeat each piece of important information from a lesson back in short worksheet answers. For example, when teaching how Rome became an empire, I’d have several short recall questions like, “Who was the main conspirator and assassin? – Brutus” or “Who did Marc Antony ally with in his war against Octavian? – Cleopatra.”

But two things happened that made me rethink everything. First, I caught students copying and pasting their worksheets into ChatGPT, gaining nothing from the exercise. Second, when they were studying for their test, I noticed that most students were creating their own study notes with this information anyway. I realized the worksheets I was giving them weren’t as valuable as I had hoped, and it was time to rethink my approach.

The Problem with Worksheets in the AI World

The safest assumption we can make today is that students will use AI to complete their work when we’re not looking. There’s little to no value left in asking students to answer simple recall questions, define words, or give one-line responses. These are too easy and too tempting to copy-paste into ChatGPT for instant answers.

AI can complete an entire lesson’s work in seconds. And even if some students don’t do this, we as teachers can never be certain which work is truly authentic. To respond to this new reality, we need to reframe how we approach worksheets altogether.

The Shift in Mindset

The goal shouldn’t be to outsmart AI, but rather,  to design work that requires human context. Real learning doesn’t happen when students simply repeat information back to us. It happens when they internalize knowledge, synthesize it, form connections, and make judgments based on their own reasoning.

That’s the kind of thinking AI can’t do for our students. If you ask students for their unique reasoning or personal perspectives, AI might assist them, but it can’t replace the human thought process behind their answer.

Design Principles for AI-Resistant Worksheets

Here are some principles I now stick to when creating worksheets:

1. Less Is More

I try to limit the number of questions to just a few per lesson. I generally divide my lessons into sections, and each section only has one or two deep questions. Fewer questions allow more time for reflection and discussion, and the responses are always richer as a result.

2. Make the Questions Student- or Context-Specific

Focus on asking students their thoughts about the content. Use words and phrases like you, your opinion, and why do you think. Ask students to reflect on the lesson content or write from an empathetic perspective. These types of questions require a more personal response than simple recall, reducing the likelihood of copy-paste answers.

3. Use Structured Creativity

I’ve created a toolkit of structured, AI-resistant tasks that I’ll write more about in the future, but here are a few examples that work especially well:

  • Newspaper headlines – Students internalize the information and rephrase it as an attention-grabbing headline in under 10 words.
  • Rank & Explain – After teaching a topic with several causes or elements, have students rank their top three and explain their reasoning. This forces evaluation and judgment.
  • Perspective journal entries or “what if” scenarios – Ask students to step into another person’s shoes and write a short journal entry from that perspective. Alternatively, ask students to predict outcomes had events happened differently. This builds empathy and requires a solid understanding of the topic.
  • Modern Mirror – Have students think of a modern situation that connects to the lesson and explain the similarity.
  • Personal connection – Ask how the content might impact their own life or connect to experiences they’ve had.

Before and After Examples

Here are some examples of traditional questions I’ve used in the past, and how I’ve redesigned them to be more AI-resistant:

Traditional QuestionAI-Resistant Redesign
What were the three main causes of the French Revolution?Rank the top three causes of the French Revolution, and explain your top choice in 3+ sentences.
List 3 similarities and 3 differences between ancient Athens and Sparta.If you had a choice of living in ancient Athens or Sparta, which city would you rather live in, and why? Write 5+ original, thoughtful sentences that include comparisons.
Explain Blitzkrieg and how the Germans used this tactic to invade France.What lessons do you think the Allied countries might have learned from Germany’s tactics when they invaded France? Explain how they might defend against similar tactics in the future (4+ sentences).

Practical Tip: Checking Authenticity Without Constant Policing

Question design matters, but supervision does too. We’d all love to tell our students, “Don’t use ChatGPT,” and just trust that they won’t – but it would be careless to do so. While AI-resistant designs help, they’re not foolproof, so here’s how I make checking for authenticity easy:

My lessons are usually divided into smaller sections, with each section addressing a particular sub-topic or skill. After I teach a section (+ related quick activity), I allow a few minutes for students to answer that section’s worksheet questions in class. During this time, I roam the room, chatting with students, guiding them, and making sure everyone is producing their own authentic answers.

This gives me a live writing sample from every student. Then, if they finish later at home, I check for consistency in voice and writing style when grading. If something feels off, I simply call the student up for a quick chat about their work. That alone sets the precedent, students quickly realize I’ll notice if they suddenly switch styles.

This method is quick, simple, and easy to build into your lesson. It keeps accountability high without making students feel constantly policed.

And there you have it. AI-resistant questions aren’t just about preventing plagiarism; they’re about encouraging authentic, individual thinking without requiring a total overhaul of your teaching style.

Happy Teaching!
Mr. G

Leave a Reply

Your email address will not be published. Required fields are marked *