Decision-Making Quality Rubric
Assessing Human-AI Integrated Decision Processes
Rubric Overview
This rubric assesses the quality of students’ decision-making processes when working with AI partners—focusing on how they integrate AI insights with human judgment in cybersecurity contexts.
Use with: Security Detective Teams, AI-Assisted Incident Response Point range: 4-16 points (4 criteria × 1-4 points each)
Assessment Criteria
Criterion 1: AI Input Integration (1-4 points)
| Score | Descriptor | Observable Behaviors |
|---|---|---|
| 4 - Advanced | Strategically integrates AI input at optimal decision points | Knows when to consult AI; synthesizes AI analysis with existing knowledge; adjusts decisions based on AI insights |
| 3 - Proficient | Consistently incorporates AI input | Regularly consults AI during decision process; uses AI analysis to inform choices |
| 2 - Developing | Inconsistent integration | Sometimes consults AI; doesn’t always incorporate insights into decisions |
| 1 - Emerging | Ignores or over-relies on AI | Either dismisses AI input entirely or accepts it without critical evaluation |
Evidence to look for:
- Timing of AI consultations (before, during, after key decisions)
- References to AI insights in decision rationale
- Balance between AI reliance and independent judgment
Criterion 2: Critical Evaluation of AI Output (1-4 points)
| Score | Descriptor | Observable Behaviors |
|---|---|---|
| 4 - Advanced | Systematically evaluates AI recommendations against multiple criteria | Questions AI reasoning; compares AI analysis to evidence; identifies potential AI errors or limitations |
| 3 - Proficient | Evaluates AI output thoughtfully | Asks follow-up questions; checks AI claims against available evidence |
| 2 - Developing | Limited evaluation | Occasionally questions AI; accepts most AI output at face value |
| 1 - Emerging | No critical evaluation | Treats AI output as authoritative; no verification attempts |
Evidence to look for:
- Follow-up questions to AI
- Comparison of AI analysis to direct evidence
- Identification of AI errors or inconsistencies
- Requests for AI to explain reasoning
Criterion 3: Human Context Application (1-4 points)
| Score | Descriptor | Observable Behaviors |
|---|---|---|
| 4 - Advanced | Expertly applies human context AI cannot access | Identifies context AI lacks; explains how context changes analysis; makes decisions AI couldn’t make |
| 3 - Proficient | Applies relevant human context | Recognizes when human knowledge matters; adds organizational/cultural context to AI analysis |
| 2 - Developing | Some context application | Occasionally adds context but doesn’t consistently recognize its importance |
| 1 - Emerging | No human context added | Relies entirely on AI analysis without adding human perspective |
Evidence to look for:
- Statements about what AI doesn’t know about the situation
- References to organizational culture, relationships, or history
- Decisions that require human judgment AI can’t replicate
Criterion 4: Decision Justification (1-4 points)
| Score | Descriptor | Observable Behaviors |
|---|---|---|
| 4 - Advanced | Articulates comprehensive justification referencing both human and AI contributions | Explains reasoning clearly; cites specific AI insights AND human factors; acknowledges trade-offs |
| 3 - Proficient | Provides clear justification | Explains reasoning with reference to AI analysis and human judgment |
| 2 - Developing | Partial justification | Provides some reasoning but may not reference both human and AI contributions |
| 1 - Emerging | No justification | Makes decisions without explaining reasoning |
Evidence to look for:
- Written or verbal explanations of decision rationale
- References to specific AI recommendations
- Acknowledgment of human factors in decisions
- Recognition of trade-offs and alternatives considered
Scoring Guide
| Total Score | Performance Level | Interpretation |
|---|---|---|
| 14-16 | Exemplary | Student demonstrates sophisticated integrated decision-making; ready for complex multi-stakeholder scenarios |
| 10-13 | Proficient | Student integrates human-AI perspectives effectively; may benefit from scenarios with greater ambiguity |
| 6-9 | Developing | Student shows emerging integration skills; needs practice with structured decision frameworks |
| 4-5 | Beginning | Student needs foundational instruction on human-AI decision integration |
Activity-Specific Application
Security Detective Teams
Focus on Criteria 1 and 2—how students integrate AI pattern recognition with their own evidence analysis.
AI-Assisted Incident Response
Focus on Criteria 3 and 4—how students apply organizational context and justify response decisions.
Instructor Notes
Key observation points:
- Decision log entries (if using)
- Group discussion contributions
- Final decision presentations
- Written reflections
Common challenges:
- Students may struggle to articulate why human context matters
- Some students over-defer to AI recommendations
- Decision justification often requires explicit prompting
Part of “True Teamwork: Building Human-AI Partnerships for Tomorrow’s Cyber Challenges” - NICE K12 2025