AI instruction failure review
Compare failed AI output against instructions and propose a fix. The submitted work creates a review brief that keeps observations tied to source material.
What the candidate submits
Failure memo, instruction check, fix priority, source examples, unknowns, revision note.
- communication artifact
- quality check
- priority rationale
- source/evidence review
- assumptions or open questions
- revision note
How this maps to the six evidence dimensions
Shows source-backed iteration, not generic AI fluency. Reviewers can inspect the submitted sources before using any observation.
Quality Rigor is supported by the quality-check portion of the artifacts: Failure memo, instruction check, fix priority, source examples, unknowns, revision note.
Ambiguity Handling is supported by assumptions, open questions, unknowns, gaps, or caveats in the submitted work.
Learning Agility is supported by the revision note and the explanation of what changed after review.
Communication Clarity is supported by the submitted update, memo, reply, handoff, guide, or summary.
Prioritization Judgment is supported by the priority rationale, sequencing choice, or next-step rationale.
Evidence Discipline is supported by source references, cited notes, logs, chart details, policy excerpts, or evidence gaps.
Questions this task can support
Which source detail would most change the next step?
Where is the evidence thin enough to ask for clarification?