Memo Critique
Students draft or receive a short legal memo, then prompt an AI tool to critique it. They evaluate the critique itself, separating useful structural feedback from overconfident, generic, inaccurate, or strategically unhelpful suggestions.
Purpose
This exercise helps students use AI as a revision partner without outsourcing their judgment.
Students begin with a legal memo draft, their own work or a sample provided by the instructor. They ask an AI tool to critique the memo, then evaluate the critique itself. The goal is not simply to accept or reject AI feedback. The goal is to decide which comments improve the memo, which comments are too generic to be useful, which comments misunderstand the law or facts, and which comments would move the draft away from the lawyer’s purpose.
The point is not to teach students that AI critique is always reliable or always useless. The point is to help students become better supervisors of feedback. AI can often identify organization problems, unclear transitions, missing headings, repetitive language, or tone issues. But it may also invent missing requirements, overstate weaknesses, flatten strategy, misunderstand the audience, or recommend changes that sound polished but make the memo less legally precise.
Learning Goals
By the end of this exercise, students should be able to:
- Prompt an AI tool to critique a legal memo for structure, clarity, completeness, tone, and legal support.
- Distinguish useful drafting feedback from generic, inaccurate, or strategically unhelpful feedback.
- Identify when an AI critique misunderstands the assignment, audience, legal standard, procedural posture, or facts.
- Evaluate whether suggested revisions would improve the memo’s legal usefulness, not just its polish.
- Decide what feedback to accept, modify, verify, or reject.
- Explain how they used AI feedback while maintaining responsibility for the final work product.
Materials
- A short legal memo draft. Students may use their own memo if appropriate for the course or setting; the instructor may instead provide a sample memo, especially for workshops, 1L instruction, asynchronous modules, or settings where students should not use live client or confidential work.
- A critique prompt or prompt template.
- A critique evaluation worksheet with categories: useful feedback, generic feedback, legally inaccurate feedback, unsupported or unverifiable feedback, strategically unhelpful feedback, and feedback requiring verification.
- Optional instructor materials: sample AI critique, model evaluation, revised memo excerpt, or debrief questions.
Setup
This exercise works best after students have already discussed AI-assisted drafting and the lawyer’s responsibility for final work product.
Before beginning, students should understand that asking AI to critique a memo is not the same thing as receiving expert legal feedback. AI may be useful as an early revision tool, for structure, clarity, headings, reader orientation, tone, and identifying possible gaps, but the student remains responsible for deciding whether the critique is legally sound, strategically appropriate, and consistent with the assignment.
The instructor should frame the exercise around supervision. Students are not only revising their memo. They are reviewing the reviewer. The central question is: what feedback should a lawyer trust, what feedback should a lawyer verify, and what feedback should a lawyer ignore?
Run of Show
Total, approximately 30–45 minutes
- Frame the task Explain that students will use AI to critique a memo draft, then evaluate the critique. The exercise is not complete when the AI gives feedback. The real work begins when students decide what to do with that feedback.
- Select or distribute the memo Students may use their own memo draft, or the instructor may provide a sample memo. If students use their own work, remind them not to upload confidential, client-identifying, or otherwise restricted material into an unapproved tool.
- Prompt the AI critique Students ask the AI tool to critique the memo. The prompt should define the audience, purpose, and desired type of feedback, structure, clarity, rule explanation, application, use of authority, tone, and missing information.
- Evaluate the critique Students review the AI feedback and sort it into categories: useful, generic, inaccurate, unsupported, strategically unhelpful, or requiring verification.
- Revise selectively Students choose two or three pieces of AI feedback to act on. They may accept the feedback, modify it, verify it first, or reject it with an explanation.
- Document judgment Students write a short note explaining which AI suggestions they used, which they rejected, and why.
- Debrief Class discussion focuses on the difference between useful critique and confident nonsense. What did AI see well? What did it miss? Where did it misunderstand the legal task? Where did the critique sound helpful but fail under closer review?
- Submission or reflection Students submit the AI critique, their evaluation of the critique, and a short explanation of what they would revise in the memo.
Student Instructions
You will work with a short legal memo draft. You may use your own memo if your instructor permits it, or you may use the sample memo provided.
First, prompt an AI tool to critique the memo. Your prompt should tell the tool:
- The audience for the memo.
- The purpose of the memo.
- The level of feedback you want.
- Whether it should focus on structure, clarity, legal analysis, use of authority, tone, or missing information.
- That it should not invent facts or assume law not provided.
After the AI produces a critique, evaluate the critique itself. Sort the feedback into these categories:
- Useful, identifies a real problem and would likely improve the memo.
- Generic, sounds reasonable but could apply to almost any memo.
- Legally inaccurate, misunderstands the law, authority, assignment, or procedural posture.
- Unsupported, makes a claim that would need to be checked before relying on it.
- Strategically unhelpful, might make the memo sound smoother but would weaken the legal analysis, client objective, or audience fit.
- Needs verification, might be right, but you would need to confirm it through the record, authorities, or assignment instructions.
Then choose two or three pieces of feedback and decide what to do with each one:
- Accept.
- Modify.
- Verify first.
- Reject.
Submit the AI critique, your evaluation, and a short explanation of which feedback you would use and which feedback you would not use.
Instructor Notes
The teaching value of this exercise comes from shifting students from “AI gave me feedback” to “I evaluated the feedback.”
Students often assume critique is safer than drafting because the AI is not producing the final work product. This exercise complicates that assumption. AI critique can improve structure and clarity, but it can also misdiagnose the problem, recommend generic fixes, or push the writer toward a memo that is more polished but less legally useful.
A strong prompt matters. Students should not ask only, “How can I improve this memo?” That kind of prompt tends to produce broad writing-center-style comments. A better prompt defines the role, audience, assignment, and review criteria.
Sample critique prompt: “Act as a supervising attorney reviewing a junior associate’s draft research memo. Focus on structure, clarity, rule explanation, application to the facts, use of authority, and tone for an internal law firm audience. Do not invent facts or assume law not included in the memo. Identify the three most important revisions needed before this could be sent to a partner. For each suggestion, explain why it matters and whether it requires legal verification.”
Common issues to watch for:
- Students may accept AI feedback because it sounds professional.
- Students may focus only on grammar and style.
- Students may overlook feedback that would change the legal analysis.
- Students may fail to distinguish structural feedback from substantive legal feedback.
- Students may accept suggestions that are inconsistent with the assignment.
- Students may reject all feedback instead of identifying what is useful.
- Students may fail to document why they accepted or rejected suggestions.
Strong student work usually shows:
- A clear critique prompt.
- Recognition that some AI feedback is genuinely useful.
- Skepticism toward generic or unsupported comments.
- Attention to legal accuracy, not just writing polish.
- Awareness of audience and purpose.
- Selective revision rather than wholesale acceptance.
- A concrete explanation of what the student would change and why.
Adaptation
Here are some suggestions on how to adapt this in other teaching contexts. The adaptations below have been limited to firm trainings and asynchronous suggestions, but there are other possible adaptations for workshops, seminars, and other contexts. Be creative!