Y12W23WR Thinking with AI vs. thinking because of AI

Design
The writing prompt

Design the specific policy your school should adopt on student use of AI tools in coursework, given both the productivity gains AI can enable and the skill-decay risks the research identifies.

1Retrieval check

Q1.What does Ethan Mollick’s research show about AI-augmented work?

  • AAI always decreases output quality
  • BOutcomes differ radically — AI as thinking partner produces large gains; AI as shortcut produces mediocre work and skill decay
  • COnly senior professionals benefit from AI
  • DAI is consistently neutral across users

Q2.Why is the skill-decay concern particularly acute for students?

  • AStudents are lazier than professionals
  • BStudents may be less able to distinguish partner-use from shortcut-use, and are still forming the underlying skills the tool substitutes for
  • CAI tools are worse for education than for work
  • DSchools ban AI so students don’t learn it
Show answer key

Q1 → B. Outcomes differ radically — AI as thinking partner produces large gains; AI as shortcut produces mediocre work and skill decay.The use-mode (partner vs. replacement) is the dominant variable in outcomes.

Q2 → B. Students may be less able to distinguish partner-use from shortcut-use, and are still forming the underlying skills the tool substitutes for.Skill formation during education is the specific concern the article flags.

2Prompt deconstruction

Stimulus
Mollick’s partner-vs-shortcut finding; the article’s worry about skill formation.
Scope
School-level policy (students, teachers, marking).
Thinking
Which tasks genuinely require unaided thinking; which benefit from partnership; how to enforce and assess.
Position
Between full prohibition and full integration — defend a specific placement.
Output
Named policy with rules, enforcement mechanism, and assessment approach.

3Position nudge

Where on the range does your proposal sit?

Pole A
Pole B

Pole AFull prohibition

Pole BFull integration

Commit to a specific point; defend it in your planner.

4Planner — design the thing, then the trade-offs

My proposal
One sentence — what the policy actually does.
Where on the range
Why this specific placement?
Covered tasks
Which tasks allow AI, in what mode (partner vs. draft-only)?
Protected tasks
Which tasks forbid AI entirely — why these, not others?
Enforcement & assessment
How is compliance checked? How is AI-assisted work marked?
Hardest objection
The strongest case against your placement — and your response.
What you won’t protect against
An honest concession — a harm your policy accepts.

5Sentence stems

  • My proposal is ___.
  • I am grounding this in [researcher]’s finding that ___.
  • The main trade-off is ___: this design gains ___ but loses ___.
  • The most predictable objection is ___, and my response is ___.
  • I would know it was working after [time] if ___.
  • What I am most likely to abandon is ___, so I will build in ___ to prevent that.

6Exemplar paragraph (not about this article)

(1) My proposal is a ‘partner-not-replacement’ policy: AI is permitted in drafting, brainstorming, and revision passes, but prohibited in the final submission of any task whose purpose is to develop the student’s own reasoning (analytical essays, mathematical proofs, structured argument pieces). (2) I am grounding this in Mollick’s finding that the partner mode produces large gains while the shortcut mode produces skill decay, and in the article’s concern that skill formation is the specific educational stake. (3) The main trade-off is complexity of marking: this design gains real pedagogical honesty but loses the simplicity of a blanket rule. (4) The most predictable objection is that students will route their ‘own reasoning’ through AI anyway, and my response is that a policy should aim at a clear cultural norm and a minimal verification step (oral follow-up on any submitted analytical piece), not at perfect prevention of cheating. (5) I would know it was working after one term if students could explain what their final submission reasoned to without access to their AI chat history. (6) What I am most likely to abandon is the oral follow-up under marking load, so I will build it into timetabled class periods rather than as additional teacher time.

What this paragraph does, move by move

  1. States the specific policy and what it covers.
  2. Grounds both in Mollick and in the article’s education-stake point.
  3. Names the marking-complexity trade-off honestly.
  4. Reframes the enforcement objection via cultural norm + minimal verification.
  5. Specifies a one-term success test.
  6. Builds timetable protection around the oral-follow-up.