StartLab
← Prompt LibraryAssessment & Feedback

Use AI for Feedback & Rubrics

Leverage AI to draft rubrics, generate calibrated exemplars, create comment banks, and structure fair, actionable feedback. Incorporate guardrails to reduce bias and support integrity-by-design.

Policy Reminder: Use of AI for Feedback and Grading

Northeastern University policy governs the use of AI tools in courses. Please review Policy 125 and follow your college or department guidance.

  • Keep a human in the loop - do not rely solely on AI to grade or determine outcomes.
  • Protect student privacy (FERPA) - avoid uploading identifiable student work to external tools.
  • Be transparent with students about how AI-assisted feedback is used in your course.
  • Verify and edit AI output, you remain responsible for accuracy, equity, and tone.
  • Document your AI use in course materials, for example the syllabus, when appropriate.

This summary is informational only. Refer to the official policy for requirements and updates.

1

Rubric from Outcomes (Criteria → Performance Levels)

Purpose: Generate a rubric aligned to outcomes with clear performance descriptors.

Act as an experienced instructional designer.

Design a rubric from stated learning outcomes for [Insert Student Level] in [Insert Course], assignment type: [Insert].

Inputs:
- Learning outcomes (3–6): [Paste]
- Assignment artifact type (e.g., report, presentation): [Insert]
- Key constraints (time, tools, word/slide limit): [Insert]
- Known common pitfalls/misconceptions: [Insert]
- Equity & accessibility considerations (language, modality): [Insert]
- AI Use Policy: [Insert] (e.g., AI may help format descriptors; not invent new criteria)

Assignment Overview:
Purpose: Translate outcomes into transparent criteria and leveled performance descriptors supporting fair, actionable feedback.

Learning Outcomes (3–5):
- Map each outcome to at least one criterion
- Distinguish adjacent performance levels with observable evidence
- Embed fairness & accessibility in wording
- Provide integrity & originality safeguards

Scenario Brief:
Students often see rubrics as opaque; explicit alignment and plain language improve self‑regulation and reduce bias.

Student Task & Deliverables:
Deliverables: Criteria list (4–6) with definitions; level descriptors (3–4 levels); evidence examples per level; integrity & originality note (proper citation, AI usage disclaimer).

Process Artifacts:
Outcome→criterion mapping table; descriptor drafting scratch notes; bias/accessibility wording checklist; evidence example ledger.

Success Criteria & Rubric Hooks:
Clear alignment, specific observable language, progressive complexity across levels, fairness/accessibility wording, integrity safeguards.

Debrief Plan (3–5 prompts):
- Where was level separation hardest?
- Which wording changes improved inclusivity?
- How will students use this to self‑assess?

Tone:
Academic, clear, student‑supportive.

Pedagogical & Ethical Requirements:
Avoid culturally biased exemplars; ensure clarity for non‑native speakers; include provenance expectations; discourage grade gaming.

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.
2

Exemplars: Good vs. Non‑Example (Annotated)

Purpose: Provide a concise exemplar and non‑example with rubric‑aligned annotations.

Act as an experienced instructional designer.

Design an exemplar vs. non‑example annotation activity for [Insert Student Level] in [Insert Course], artifact type: [Insert].

Inputs:
- Rubric criteria & level descriptors: [Paste]
- Common strong features to highlight: [Insert]
- Common weak patterns/pitfalls: [Insert]
- Accessibility & inclusivity notes (jargon to minimize): [Insert]
- AI Use Policy: [Insert] (e.g., AI may help paraphrase annotations; not fabricate data)

Assignment Overview:
Purpose: Provide one strong exemplar and one contrasting non‑example with criterion‑linked annotations to build evaluative judgment.

Learning Outcomes (3–5):
- Differentiate quality features vs. superficial polish
- Link annotations directly to rubric criteria
- Identify and articulate specific improvement actions
- Practice respectful, non‑stigmatizing critique language

Scenario Brief:
Students may overfocus on surface traits; annotated contrast clarifies underlying structural quality and growth focus.

Student Task & Deliverables:
Deliverables: Template skeleton; Annotated Good Example (inline markers + summary table); Annotated Non‑Example (issues list + improvement suggestions); 3 quick self‑check questions + answer key.

Process Artifacts:
Annotation ledger (criterion→text span→rationale); issue classification list; improvement suggestion table.

Success Criteria & Rubric Hooks:
Direct criterion linkage, specificity of rationales, constructive tone, clarity of improvement suggestions, avoidance of bias.

Debrief Plan (3–5 prompts):
- Which annotation most clarified a hidden quality dimension?
- What improvement suggestion felt most actionable?
- How will students use this before drafting?

Tone:
Academic, supportive, growth‑oriented.

Pedagogical & Ethical Requirements:
Non‑stigmatizing language, inclusivity, transparency of limitations, integrity of exemplar provenance.

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.
3

Calibration Set Builder (Anchor Papers/Artifacts)

Purpose: Create 3–5 anchor artifacts with rubric‑mapped rationales for calibration.

Act as an experienced instructional designer.

Design a calibration anchor set (3–5 artifacts) for [Insert Student Level] in [Insert Course], artifact type: [Insert].

Inputs:
- Rubric criteria & level descriptors: [Paste]
- Distribution of expected performance levels: [Insert]
- Common scorer disagreement areas: [Insert]
- Bias/ambiguity risk factors (language, topic, formatting): [Insert]
- AI Use Policy: [Insert] (e.g., AI may assist formatting tables; not alter scores)

Assignment Overview:
Purpose: Provide anchor artifacts with transparent scoring rationales to calibrate evaluators and reduce inconsistency/bias.

Learning Outcomes (3–5):
- Apply rubric consistently across varied quality levels
- Articulate criterion‑specific rationales
- Detect and mitigate potential bias sources
- Refine shared scoring norms via discussion

Scenario Brief:
Without anchors, early scoring drifts and hidden biases propagate; structured rationale stabilizes norms before high‑stakes evaluation.

Student Task & Deliverables:
Deliverables: Anchor Artifacts A–E (excerpt or condensed version); Score Table (criterion→level per artifact); Rationale Notes (criterion→evidence→justification); Bias/Ambiguity Flag List (if any); Discussion Prompts (3).

Process Artifacts:
Rationale drafting worksheet, discrepancy log (initial vs consensus), bias flag ledger.

Success Criteria & Rubric Hooks:
Consistent criterion application, clear evidence citation, proactive bias flagging, convergence toward consensus.

Debrief Plan (3–5 prompts):
- Which criterion produced most disagreement and why?
- What evidence clarified scoring alignment?
- How will anchors inform future revisions to rubric wording?

Tone:
Academic, transparent, norm‑building.

Pedagogical & Ethical Requirements:
Fairness, inclusivity, integrity (no fabricated evidence), clear provenance of anchor artifacts.

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.
4

Comment Bank Generator (Criterion‑Aligned)

Purpose: Draft supportive comments mapped to rubric criteria and performance levels.

Act as an experienced instructional designer.

Design a criterion‑aligned comment bank for [Insert Student Level] in [Insert Course], artifact type: [Insert].

Inputs:
- Rubric criteria & levels: [Paste]
- Common pitfalls per criterion: [Insert]
- Desired strengths to highlight: [Insert]
- Tone guardrails (avoid vague praise, ensure specificity): [Insert]
- Accessibility & inclusivity considerations: [Insert]
- AI Use Policy: [Insert] (e.g., AI may draft initial phrasing; human reviews for bias)

Assignment Overview:
Purpose: Provide reusable, specific, fair comments mapped to each criterion×performance level supporting actionable student revision.

Learning Outcomes (3–5):
- Produce comments referencing observable evidence
- Differentiate constructive improvement vs. reinforcement of strengths
- Mitigate bias and vague language
- Support student self‑regulation and revision planning

Scenario Brief:
Unguided comments drift toward generic praise or harshness; structured bank ensures consistent, equitable guidance.

Student Task & Deliverables:
Comment Tables: Criterion rows × Level columns with 2–3 comments each (strength + improvement). Usage Tips: How to select/adapt comments responsibly. Integrity Note: Cite sources, disclose AI assistance.

Process Artifacts:
Draft comment worksheet, bias/inclusivity checklist, adaptation log (selected comment→final delivered).

Success Criteria & Rubric Hooks:
Specificity, criterion alignment, balanced tone, actionable improvement step, inclusivity.

Debrief Plan (3–5 prompts):
- Which criterion was hardest to phrase specifically?
- What patterns risked bias?
- How will adaptation logs improve feedback quality?

Tone:
Academic, supportive, equity‑minded.

Pedagogical & Ethical Requirements:
Avoid stereotyping, respect privacy, disclose AI assistance, encourage authentic personalization.

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.
5

AI Feedback Guardrails (Prompt + Policy)

Purpose: Create a safe, course‑specific AI feedback prompt with do/don’t and attribution rules.

Act as an experienced instructional designer.

Design AI formative feedback guardrails for [Insert Student Level] in [Insert Course], allowed use scope: [Insert].

Inputs:
- Course rubric highlights: [Paste]
- Permitted AI assistance types: [Insert]
- Prohibited AI actions (e.g., grading, generating final artifact): [Insert]
- Privacy & data handling constraints: [Insert]
- Academic integrity expectations (citation/logging): [Insert]

Assignment Overview:
Purpose: Provide a safe AI prompt + policy ensuring feedback is supportive, bias‑aware, and integrity‑preserving.

Learning Outcomes (3–5):
- Use AI to obtain formative, criterion‑aligned suggestions
- Apply guardrails to prevent overreliance or plagiarism
- Log and attribute AI assistance transparently
- Evaluate AI feedback quality vs rubric criteria

Scenario Brief:
Students risk copying AI rewrites or exposing private data; explicit policy structures safe engagement.

Student Task & Deliverables:
Sections: AI Feedback Prompt (structured inputs & required context); Do/Don’t Table; Privacy & Ethics Guidelines; Attribution & Logging Format; 2–3 Example Interactions (good vs. unsafe).

Process Artifacts:
AI usage log (prompt→response→action taken), attribution statement template, safeguard checklist.

Success Criteria & Rubric Hooks:
Criterion alignment, transparency, responsible scope, privacy compliance, original student authorship.

Debrief Plan (3–5 prompts):
- Which guardrail most prevents misuse?
- How will logging change reflection quality?
- What improvements might refine safe scope?

Tone:
Academic, responsible, clarity‑focused.

Pedagogical & Ethical Requirements:
No undisclosed AI generation, respect privacy, discourage wholesale substitution, promote reflective integration.

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.
6

Self & Peer Assessment Scaffold

Purpose: Provide checklists and prompts to support fair self/peer review.

Act as an experienced instructional designer.

Design a self & peer assessment scaffold for [Insert Student Level] in [Insert Course], artifact: [Insert].

Inputs:
- Rubric criteria & descriptors: [Paste]
- Common peer review pitfalls (halo, harshness, vagueness): [Insert]
- Inclusivity & civility guidelines: [Insert]
- AI Use Policy: [Insert] (e.g., AI may help format comments; must disclose)

Assignment Overview:
Purpose: Support fair, criterion‑aligned self & peer review that generates actionable revision insights and mitigates bias.

Learning Outcomes (3–5):
- Apply criteria consistently to own and peers' work
- Provide balanced, specific feedback (strength + improvement)
- Detect and mitigate bias tendencies
- Reflect on feedback incorporation strategy

Scenario Brief:
Unguided peer review can drift toward unhelpful praise or vague critique; structured checklists and reflection anchor quality.

Student Task & Deliverables:
Self Checklist (criterion evidence ticks); Peer Checklist (observed strengths, prioritized improvements); Feedback Prompts (3–5 targeted criterion questions); Reflection Form (revision intentions + anticipated impact); Integrity Note (original work affirmation + AI usage disclosure).

Process Artifacts:
Completed self checklist, peer feedback forms, revision intention summary, AI usage log (if any).

Success Criteria & Rubric Hooks:
Criterion specificity, respectful tone, improvement feasibility, bias awareness, reflective revision planning.

Debrief Plan (3–5 prompts):
- Which feedback item will most improve quality?
- Where did bias mitigation change your phrasing?
- What revision plan risks remain?

Tone:
Academic, supportive, civility‑focused.

Pedagogical & Ethical Requirements:
No stigmatizing language, balanced framing, transparency of sources, encourage growth mindset.

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.
7

Feedback → Revision Plan Mapper

Purpose: Turn feedback into a concrete revision log with priorities and evidence.

Act as an experienced instructional designer.

Design a feedback → revision plan mapping activity for [Insert Student Level] in [Insert Course], artifact: [Insert].

Inputs:
- Feedback summary (grouped by criterion): [Paste]
- Rubric criteria & performance goals: [Insert]
- Deadlines / time constraints: [Insert]
- Available support resources (tutoring, office hours): [Insert]
- AI Use Policy: [Insert] (e.g., AI may suggest rewrite options; student chooses & cites)

Assignment Overview:
Purpose: Translate received feedback into prioritized revision actions with evidence tracking and integrity safeguards.

Learning Outcomes (3–5):
- Classify feedback by criterion & type (concept, structure, clarity)
- Prioritize changes by impact & effort
- Specify revision actions with evidence reference
- Maintain transparent authorship & citation

Scenario Brief:
Students often act on easy fixes first; structured prioritization ensures impactful improvements and tracks evidence of change.

Student Task & Deliverables:
Plan Table: Issue (feedback excerpt) → Planned Change → Evidence Source (section/page/data) → Priority (High/Med/Low + rationale) → Status. Success Criteria List (4–6). Integrity Note (original work + citation plan).

Process Artifacts:
Feedback categorization ledger, prioritization rationale notes, revision status updates log.

Success Criteria & Rubric Hooks:
Impactful prioritization, clear action descriptions, evidence linkage, progress transparency, integrity adherence.

Debrief Plan (3–5 prompts):
- Which planned change yields largest quality gain?
- What low‑effort change was deprioritized and why?
- How will you validate improvement after revision?

Tone:
Academic, constructive, prioritization‑focused.

Pedagogical & Ethical Requirements:
Avoid punitive framing, emphasize growth, cite all external ideas, disclose AI assistance.

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.
8

Rubric → Student Checklist (Plain Language)

Purpose: Translate rubric into a plain‑language checklist for students.

Act as an experienced instructional designer.

Design a plain‑language student checklist derived from rubric criteria for [Insert Student Level] in [Insert Course], artifact: [Insert].

Inputs:
- Rubric criteria & key level distinctions: [Paste]
- Common student misunderstandings: [Insert]
- Accessibility & language simplification needs: [Insert]
- Integrity expectations (citation/originality): [Insert]
- AI Use Policy: [Insert] (e.g., AI may help simplify wording; verified by instructor)

Assignment Overview:
Purpose: Convert rubric into a concise, student‑friendly checklist aiding self‑assessment before submission.

Learning Outcomes (3–5):
- Translate criteria into actionable self‑check items
- Identify gaps prior to submission
- Apply integrity & citation expectations proactively
- Reduce ambiguity in expectations

Scenario Brief:
Students struggle to decode formal rubric language; accessible phrasing supports equitable success.

Student Task & Deliverables:
Checklist (5–8 items plain language); Mapping Table (Checklist Item → Original Criterion); Quick Self‑Check (3 yes/no + interpret next action); Integrity Reminder (authorship, sources, AI disclosure).

Process Artifacts:
Draft simplification notes, mapping table, self‑check responses.

Success Criteria & Rubric Hooks:
Clarity, direct criterion linkage, actionable phrasing, integrity reinforcement, accessibility.

Debrief Plan (3–5 prompts):
- Which checklist item most often triggered revision?
- What phrasing change improved comprehension?
- How will students maintain integrity while using the checklist?

Tone:
Academic, clear, student‑centered.

Pedagogical & Ethical Requirements:
Inclusive, non‑stigmatizing language, transparency of expectations, reinforcement of proper citation & originality.

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.
9

Bias‑Aware Moderation Checklist

Purpose: Provide checks to detect/mitigate bias in feedback and scoring.

Act as an experienced instructional designer.

Design a bias‑aware moderation checklist for rubric‑based feedback/scoring in [Insert Course], level: [Insert Student Level].

Inputs:
- Rubric criteria & descriptors: [Paste]
- Sample feedback comments (anonymized): [Insert]
- Known bias risks (language, assumption, stereotype): [Insert]
- Scoring consistency pain points: [Insert]
- AI Use Policy: [Insert] (e.g., AI may flag potential bias; human confirms)

Assignment Overview:
Purpose: Provide structured checks, examples, and remediation steps to reduce bias and improve scoring consistency.

Learning Outcomes (3–5):
- Detect linguistic & structural bias in comments
- Apply consistency checks across similar artifacts
- Remediate biased phrasing into neutral, specific language
- Document moderation adjustments transparently

Scenario Brief:
Unmoderated feedback can reinforce inequities; systematic checks enhance fairness and student trust.

Student Task & Deliverables:
Checklist Categories: Language Neutrality, Criterion Consistency, Evidence Citation, Tone Balance. Examples (biased vs. revised). Remediation Steps (identify→revise→log). Integrity Note (transparency & logging).

Process Artifacts:
Moderation log (original→revised comment), bias flag ledger, consistency comparison table.

Success Criteria & Rubric Hooks:
Reduced biased phrasing, consistent scoring, clear evidence referencing, transparent revisions.

Debrief Plan (3–5 prompts):
- Which bias category appeared most?
- What remediation pattern improved clarity?
- How will logs inform future training?

Tone:
Academic, responsible, fairness‑focused.

Pedagogical & Ethical Requirements:
Inclusive language, avoid stereotypes, transparent correction trail, respect privacy.

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.
10

Rubric Alignment Audit (Objectives ↔ Criteria ↔ Evidence)

Purpose: Audit alignment among objectives, criteria, and evidence artifacts.

Act as an experienced instructional designer.

Design a rubric alignment audit for [Insert Student Level] in [Insert Course], assignment: [Insert].

Inputs:
- Learning objectives list: [Paste]
- Current rubric criteria & descriptors: [Insert]
- Sample student artifact excerpts (anonymized): [Insert]
- Known misalignment complaints (student/faculty): [Insert]
- Equity & validity concerns: [Insert]
- AI Use Policy: [Insert] (e.g., AI may assist table formatting; not invent objectives)

Assignment Overview:
Purpose: Map objectives ↔ criteria ↔ evidence to surface gaps, redundancies, and fairness issues for targeted rubric refinement.

Learning Outcomes (3–5):
- Identify coverage/gap patterns across objectives
- Distinguish redundant or overlapping criteria
- Evaluate evidence sufficiency & representativeness
- Recommend precise rubric refinements

Scenario Brief:
Misaligned rubrics erode trust and distort focus; transparent audit guides iterative improvement.

Student Task & Deliverables:
Alignment Table (Objective → Associated Criteria → Evidence Examples → Coverage Rating). Gap Analysis (missing or under‑represented objectives). Redundancy/Overlap Notes. Recommendations (specific wording or structural changes).

Process Artifacts:
Objective–criterion mapping worksheet, evidence sampling log, revision recommendation ledger.

Success Criteria & Rubric Hooks:
Comprehensive mapping, clear gap identification, actionable recommendations, fairness/validity awareness.

Debrief Plan (3–5 prompts):
- Which objective lacked strongest evidence?
- What criterion overlapped most?
- How will recommended changes improve fairness?

Tone:
Academic, structured, improvement‑focused.

Pedagogical & Ethical Requirements:
Fair coverage, clarity for diverse learners, integrity of evidence selection (no cherry‑picking).

Before producing the final version, ask me three clarification questions to confirm understanding. Then summarize my answers back to me in 3 bullets. Only after that, present the final output in clearly labeled sections.