A grading workflow: a rubric created with AI, AI-generated feedback for 3 sample student essays, and a calibration process to make sure AI feedback matches your standards.
What Ethical AI Grading Looks Like
Let's be direct: AI should not assign final grades. That's a professional and ethical responsibility that belongs to the teacher. What AI can do is generate detailed, specific feedback at a scale that would take you hours to produce manually.
The distinction:
AI does:
→ Draft specific, actionable feedback per criterion
→ Identify patterns across student work
→ Suggest scores based on your rubric (for your review)
→ Generate personalized next steps
Teacher does:
→ Review and adjust AI feedback before sharing
→ Apply professional judgment to edge cases
→ Make final grade decisions
→ Have follow-up conversations with studentsThink of AI as a highly capable teaching assistant who reads every paper and writes a feedback draft. You review and approve. This is both ethical and practical.
Build a Rubric with AI
Good AI feedback depends on a good rubric. AI can help you build one faster and more completely than you would alone.
Create a 4-point rubric for the following assignment:
Assignment: [describe the assignment in 2-3 sentences]
Grade level: [grade]
Key criteria: [list 3-5 things you're assessing]
For each criterion, provide descriptors for:
4 (Exceeds), 3 (Meets), 2 (Approaching), 1 (Beginning)
Make descriptors specific and observable,
not vague (avoid "good" and "poor").Once you have the rubric, save it — you'll paste it into every AI feedback prompt for that assignment. The rubric is what calibrates AI feedback to your actual standards.
Generate and Review Feedback
With a rubric in hand, the feedback prompt is straightforward:
You are helping a teacher provide feedback on student work.
Use the rubric below to guide your feedback.
Be specific — quote or reference the student's own words.
Be constructive — explain how to improve, not just what's wrong.
Be age-appropriate for [grade level] students.
Rubric:
[paste your rubric here]
Student work:
[paste student submission here]
Provide:
1. Suggested score per criterion with 1-2 sentence justification
2. Two specific strengths
3. Two specific areas to improve with concrete guidance
4. One personalized next stepCalibration step: Before using AI feedback on student work, grade 3 papers yourself first. Then run those same papers through AI and compare. Where does AI diverge from your judgment? Adjust the prompt or rubric to close the gap. This calibration turns AI feedback from "generic" to "aligned with my standards."
Always review before sharing. AI occasionally misreads student intent or misses context you'd catch. Read every feedback draft before sending it to students. This takes 2-3 minutes per student instead of 10-15 — still a major time saving.
What You Learned Today
- The ethical framework for AI grading: AI drafts, teacher decides
- How to build a specific, observable rubric with AI
- The feedback prompt structure: rubric + student work + specific output format
- Why calibration (comparing AI grades to yours) is the critical step before deploying
Go Further on Your Own
- Take 5 student papers you've already graded and run them through AI using your rubric. How closely does AI match your grades? Where does it diverge?
- Ask AI to generate a 'common patterns' summary after analyzing 10 student papers: what misconceptions appear most often? Use this to plan reteaching.
- Build a 'feedback bank' — ask AI to generate 20 specific feedback phrases for your most common rubric criteria. Use these to write feedback faster yourself.
Nice work. Keep going.
Day 3 is ready when you are.
Continue to Day 3Want live instruction and hands-on projects? Join the AI bootcamp — 3 days, 5 cities.