Student learning · Grades 6–8 · 45 minutes
Did the AI Do the Thinking?
Students do one thinking task by hand and a matching one with AI, then take a no-notes challenge that reveals what they actually learned, surfacing cognitive offloading, when a tool helps you think and when it thinks instead of you.
AI can hand you an answer in seconds, which feels like progress but can quietly do the thinking you were supposed to do yourself. This lesson makes that trade visible. Pairs complete two short, similar thinking tasks: one by hand and one where an approved AI does most of the work. Right after, each student faces a short no-notes, no-AI challenge on both topics, and they discover they can usually explain and extend the one they struggled through, but go blank on the one the AI handled. The class names this pattern as cognitive offloading and sorts when offloading is smart (a calculator for arithmetic you already understand) from when it robs you of learning (letting AI reason through something you were trying to learn). The stance is balanced: AI is not banned and not magic. The question students leave with is the one the ELEs ask, what evidence shows the tool improved my learning, not just my output?
In this packet
- Handout A: Two Tasks (worksheet)
- Handout B: The No-Notes Challenge (worksheet)
- Handout C: Smart or Sorry Offload? (card sort)
- Facilitator key (last page)
TCEA ELE indicators
- AI4.1 I know how to promote critical thinking skills in relation to AI-generated information.
- S4.1 I know how to take responsibility for my own learning.
- S4.3 I am aware of my learning strengths and areas for improvement.
- T5.1 I know how to design lessons that promote higher-order thinking skills.
Full facilitator guide: mglearn.github.io/eles/activities/did-the-ai-do-the-thinking.html
Handout A for Did the AI Do the Thinking?
Two Tasks
Do Task 1 with only your brains, notes, and partner. Do Task 2 by letting the approved AI solve and explain it, then copy its answer. Do not improve the AI's work, just let it do the thinking this time.
TASK 1 (by hand): A classmate argues, "Our team lost after we changed our warm-up, so the new warm-up made us lose." Explain, step by step, why this reasoning might be wrong and what else could explain the loss.
TASK 2 (with AI): Ask the AI to explain why this reasoning may be flawed: "Sales went up the month after the new sign went up, so the sign caused the sales." Copy the AI's explanation here.
Quick gut check: which task felt like more effort? Which felt faster?
Handout B for Did the AI Do the Thinking?
The No-Notes Challenge
No notes. No AI. No partner. Just you. This is not graded, it is an experiment on what stuck.
In your own words, explain the thinking error in BOTH arguments from Task 1 and Task 2. Which one is easier to explain now?
NEW ONE: "It rained the day after I washed my car, so washing my car caused the rain." Explain why this reasoning is wrong.
Be honest: which task do you understand better right now, the one you did by hand or the one the AI did? Why?
MY THINKING RULE, part 1: one thing today's experiment showed me about my own learning:
MY THINKING RULE, part 2: my rule for using AI when I'm trying to LEARN something new:
MY THINKING RULE, part 3: one kind of thinking I never want to offload, and why:
Handout C for Did the AI Do the Thinking?
Smart or Sorry Offload?
Cognitive offloading means handing a mental task to a tool. Sort each one: is it a SMART offload (fine to hand off) or a SORRY offload (you lose learning you needed)? Tell why.
Smart offload
Sorry offload
Handout C for Did the AI Do the Thinking?
Smart or Sorry Offload?: cards to sort
Cut apart the strips. Sort each one onto the mat and be ready to explain why.
Using a calculator for long division you already know how to do
Letting AI write your essay about a book you were supposed to read
Using a map app to navigate a city once you can already read a map
Having AI solve the exact kind of math problem you have a test on tomorrow
Asking AI to check your finished reasoning for holes
Letting AI pick your argument and evidence so you don't have to decide
Using spell-check on a lab report
Having AI summarize a chapter so you can skip reading it for discussion
Asking AI to translate a word you don't know in a language you're fluent in
Letting AI do your coding assignment meant to teach you to code
Facilitator only for Did the AI Do the Thinking?
Answer key and notes
Handout C: Smart or Sorry Offload?
| Item | Sort | Why |
|---|---|---|
| Using a calculator for long division you already know how to do | Smart offload | You mastered the skill; the tool just saves time. |
| Letting AI write your essay about a book you were supposed to read | Sorry offload | You skip the reading and thinking the assignment was built to grow. |
| Using a map app to navigate a city once you can already read a map | Smart offload | The underlying skill is there; the tool is convenience. |
| Having AI solve the exact kind of math problem you have a test on tomorrow | Sorry offload | You bypass the practice that builds the skill you'll be tested on. |
| Asking AI to check your finished reasoning for holes | Smart offload | You did the thinking first; AI is a second set of eyes. |
| Letting AI pick your argument and evidence so you don't have to decide | Sorry offload | Deciding and defending a claim is the thinking you're learning to do. |
| Using spell-check on a lab report | Smart offload | Spelling isn't the skill the report is assessing. |
| Having AI summarize a chapter so you can skip reading it for discussion | Sorry offload | You lose the close reading and can't really join the discussion. |
| Asking AI to translate a word you don't know in a language you're fluent in | Smart offload | A quick lookup of one word, not the core task. |
| Letting AI do your coding assignment meant to teach you to code | Sorry offload | The struggle of writing and debugging is how you learn to code. |