Job to be done
Use AI in an assessment without hiding its role or weakening the evidence of learning.
Evidence state: Editorial reference protocol; no community outcome is claimed.
Baseline
- Identify the learning objective
- Record the current assessment rubric
- Document accessibility and privacy constraints
Method
- Label AI-permitted stages
- Add process evidence requirements
- Test the rubric on sample work
- Collect learner feedback without personal data
Evaluation
- Objective alignment
- Process evidence quality
- Accessibility
- Disclosure compliance
Limitations
- Institutional policy and applicable law take precedence.
- The protocol does not determine academic misconduct.