AI Curriculum Currency and Graduate Competency Gap Audit

Audit whether an academic programme develops current, durable and assessable AI-related graduate competencies using supplied curriculum evidence and dated external sources.

## Audit inputs

Programme outcomes, module specifications and assessment evidence:
{{programme_and_assessment_evidence}}

Learner profile, discipline, graduate destinations and institutional mission:
{{programme_context}}

Primary standards, credible research and labour evidence:
{{external_evidence}}

Review scope, resource constraints and accountable owners:
{{review_constraints_and_owners}}

## Evidence discipline

- Separate current curriculum evidence, direct external evidence, stakeholder claim, reviewer inference, assumption, conflict, missing information and unresolved uncertainty.
- Record source title, author or issuing body, jurisdiction or population, version or status, publication date, access date, method, direct URL or identifier and applicability for every consequential external claim. Check whether a standard or framework has been withdrawn or superseded. Never invent a standard, citation, labour forecast or employer requirement.
- Prefer primary standards, competency frameworks, systematic evidence and transparent labour data over vendor marketing or isolated job advertisements.
- Treat job-posting counts, employer surveys, occupational data and forecasts as different evidence types. Record their coverage, sampling and limitations; do not infer durable labour demand from vacancy frequency alone or claim that a curriculum change will improve graduate employment without suitable outcome evidence.
- Distinguish durable competencies, such as evidence judgment or data stewardship, from product-specific operation that may decay quickly.
- Do not infer teaching or learner attainment from a module title. Require teaching, practice and assessment evidence.
- Preserve academic governance: the review proposes evidence-backed changes, while authorized programme bodies approve outcomes, assessments and curriculum.
- For consequential external evidence, record the original-source URL or DOI, publication date and access date, then require manual verification in the original source. Do not treat a search summary or generated citation as evidence.
- Exclude learner records and sensitive attainment data unless the review explicitly requires aggregated evidence and the institution has approved the handling method.

## Audit method

1. Define the graduate roles, discipline context, time horizon and decision the audit will inform.
2. Normalize external evidence into competency statements with source version or status, publication and access dates, population, method, direct locator, applicability, confidence, limitation and expected shelf life. Keep observed demand, stakeholder preference and forecast demand distinct.
3. Map each programme outcome to modules, teaching activity, practice opportunity, assessment, performance standard and progression stage.
4. Test coverage for technical use, critical evaluation, evidence and data literacy, domain application, ethics/safety, accessibility, collaboration and accountable decision-making only where relevant.
5. Identify underassessment, duplication, premature advanced content, progression breaks, obsolete tool-specific content and unsupported claims of coverage.
6. Challenge each proposed addition: why it matters, supporting evidence, likely durability, prerequisites, opportunity cost and how attainment would be assessed.
7. Compare the smallest viable options: refresh examples, revise an assessment, strengthen progression, add a module element, retire content, or gather better evidence.
8. Create an evidence-refresh schedule based on source volatility rather than an arbitrary annual rewrite.

## Output contract: Curriculum Currency and Competency Dossier

Return:

1. **Scope and evidence register**: programme context, decision horizon, source title and issuing body, version or status, publication/access date, direct URL or identifier, jurisdiction/population, method, applicability, supersession check and reliability limit.
2. **Competency architecture**: competency, durable/vendor-specific classification, rationale, prerequisites and observable attainment.
3. **Curriculum-to-assessment map**: outcome, module, teaching, practice, assessment, standard, progression and evidence status.
4. **Gap and redundancy register**: issue, evidence, consequence, affected learners, priority and uncertainty.
5. **Change options**: smallest intervention, intended learning outcome, assessment evidence, workload/resource effect, dependency and owner.
6. **Vendor-fashion challenge**: proposed content, supporting source, durability risk, portable alternative and decision.
7. **Review recommendation**: `Maintain`, `Targeted refresh`, `Programme redesign evidence phase`, or `Insufficient evidence`, with rationale.
8. **Currency plan**: volatile evidence to monitor, trigger, cadence, source owner and next formal review.

## Verification and completion

Complete only when external evidence is directly traceable, dated, checked for current or superseded status and qualified for method and applicability; claimed curriculum coverage has teaching and assessment support; durable and vendor-specific competencies are distinguished; changes have assessable outcomes and opportunity costs; and approval authority remains explicit.

For each completion condition, record the expected curriculum or external evidence, the actual supplied evidence, `Met`, `Not met`, or `Blocked` status, and the unresolved programme-owner action.

If programme documents or credible external evidence are incomplete, return a provisional map and prioritized evidence plan. Refuse requests to fabricate labour demand, declare a programme compliant or present proposed changes as approved.