Rubrics and scoring AI outputs

Rubrics and scoring AI outputs

Why it matters: Rubrics turn subjective “good” into a consistent scoring system.

Goals

• Write a rubric with criteria.
• Score outputs reliably.
• Use the rubric to guide revisions.

Key definitions

• Rubric: A scoring guide with criteria and levels (e.g., 1–5).
• Criterion: One dimension of quality (accuracy, clarity, completeness, tone).
• Anchor example: A small example that represents a score level.

Workflow (step-by-step)

1. Pick 4–6 criteria and define them.
2. Create 3 score levels (poor/okay/great) with anchors.
3. Score 3 outputs and discuss discrepancies.
4. Revise prompts to target the lowest-scoring criterion.
5. Re-run scoring to verify improvement.

Example (good vs bad)

✅ Good: You can score outputs consistently and improve prompts based on rubric feedback.

❌ Bad: Quality depends on who reviews it and what mood they’re in.

Checklist

• Criteria defined.
• Anchors exist.
• Scores are repeatable.
• Rubric informs revisions.

Metrics / criteria

• Inter-rater agreement improves; rubric scores trend up.
• Measurable: You can say whether the output is correct/complete (pass/fail or a score).
• Effort: You can produce the result in 15–30 minutes using the workflow.
• Clarity: Another person can run your prompt and get a similar outcome.

Common mistakes + fixes

• Pitfall: Rubric too vague. Fix: Add concrete anchors and examples per level.
• Pitfall: Vague instructions. Fix: Add constraints, examples, and a success checklist.
• Pitfall: One-shot prompting. Fix: Iterate: draft → critique → revise → verify.
• Pitfall: No validation. Fix: Add QA steps, tests, and explicit “what counts as done”.

Student tasks

1. Apply the workflow to one real work task related to rubrics and scoring ai outputs.
2. Create a small artifact: prompt, checklist, rubric, table, draft, or decision note.
3. Run one quality check and record what changed after the check.
4. Write a 3-sentence reflection: what worked, what failed, and what you will reuse.

Useful external sources

• OpenAI prompt engineering guide: https://platform.openai.com/docs/guides/prompt-engineering — Prompt structure, iteration, constraints, and evaluation habits.
• NIST AI Risk Management Framework: https://www.nist.gov/itl/ai-risk-management-framework — Risk framing, governance language, and verification mindset.
• OWASP Top 10 for Large Language Model Applications: https://owasp.org/www-project-top-10-for-large-language-model-applications/ — Practical AI safety risks such as prompt injection, leakage, and insecure output handling.
• Google People + AI Guidebook: https://pair.withgoogle.com/guidebook/ — Human-centered AI design, user control, feedback, and failure modes.

Bibliography

• OpenAI prompt engineering guide. https://platform.openai.com/docs/guides/prompt-engineering
• NIST AI Risk Management Framework. https://www.nist.gov/itl/ai-risk-management-framework
• OWASP Top 10 for Large Language Model Applications. https://owasp.org/www-project-top-10-for-large-language-model-applications/
• Google People + AI Guidebook. https://pair.withgoogle.com/guidebook/