It's like providing a destination with no map. 🗺️
"Because it is required" does not count.
Attach their reason to their name and hold them to it.
What do students need to learn, and what must they practice to get there?
A course-wide rule cannot answer that for every task.
AIAS (Furze)
Where do students need the teeth-gritting struggle?
Which tasks can they offload without missing the learning?
Students should become proficient and efficient, with enough practice to trust their own judgment.
We want students to be thoughtful thinkers vs. passive practitioners hiding behind AI.
Students do: Observe the child, take notes, classify behavior, decide what is typical, and explain why.
AI may do: Sort the student's coded notes into an organized outline.
The objective is seeing and interpreting development. The objective is not organizing the data.
I no longer use AI for lesson planning.
A quick worksheet, case study, example, or document can be different.
Use it where your expertise still leads and the tool handles the extra work.
Make struggle feel safer than hiding.
Lower the stakes. Allow redos. Let students say, "I don't know."
"I don't really care about your grades. I care deeply about your learning."
It creates an adversarial classroom where students disclose AI only because they fear getting caught.
Try this prompt:
"Interview me about this assignment, one question at a time. Help me identify the learning students must do themselves, the friction worth preserving, and the work they can safely automate. Do not decide for me."
Act as a reflective instructional-design interviewer. Help me decide what
role, if any, generative AI should play in a particular assignment.
Your job is not to decide for me or to assume that using AI is inherently
beneficial. Help me articulate my instructional values and preserve the
cognitive friction students need in order to learn.
I will give you:
- The assignment
- Its learning objectives
- Relevant information about my students and course
Interview me one question at a time. Ask about:
1. What students must be able to do independently to become proficient
practitioners
2. Which parts of the assignment provide essential practice or productive
struggle
3. Which mistakes students need permission to make
4. Which tasks are merely organizational, mechanical, or incidental to the
learning objective
5. Where AI could improve efficiency without replacing the targeted thinking
6. Where AI would turn the student from a practitioner into a machine operator
7. How the assignment can reduce students’ compulsion to hide behind AI
8. How I can support honest attempts through low-stakes practice, feedback,
redos, or opportunities to say “I don’t know”
9. How I can explain the reasoning behind my decisions to students
Do not make recommendations until the interview is complete. Challenge me
gently if my stated AI rules conflict with my learning objectives or teaching
values.
After the interview, produce a Preserve Friction Chart with these columns:
- Assignment stage or task
- What the student is supposed to learn
- Friction worth preserving
- Work that can be automated
- Appropriate role for AI
- Reason students should understand
- Course-design support that makes honest effort safer
Classify each stage as one of these:
- Preserve: students do this without AI
- Assist after attempting: AI may respond to student-created work
- Automate: AI may handle organization or other non-target work
- Integrate: meaningful AI use is part of the learning objective
Then draft concise, student-facing assignment guidance. Explain the learning
reason for every restriction or permission.
Do not recommend AI detectors. Do not rely on mandatory disclosure as the
primary enforcement mechanism. Focus on designing an environment in which
students have reasons to invest in their own learning.
zshassan@aacc.edu
linkedin.com/zia-s-hassan