When Legacy Curriculum Meets Artificial Intelligence: A Theoretical Framework for Institutional Renewal in Career and Technical Education

Career and technical education does not become current by adding generative AI assignments to courses it already offers. Wang argues that AI compresses the useful life of a curriculum, which moves the governing question from whether a course can still be scheduled to whether its inclusion can be justified by contemporary occupational knowledge, degree-level rigor, and external evidence. The Adaptive CTE Renewal Framework answers that with four foundations, five dimensions, six leadership tests, and three institutional states, and it replaces periodic program review with continuous curriculum intelligence.

Key Findings

• Administrative continuity is not curricular relevance. A program can enroll students, schedule courses, and award degrees while losing alignment with the occupation it claims to prepare people for. Functioning proves a program operates, not that its curriculum is current.
• AI compresses the useful life of curriculum. Because AI redistributes tasks rather than eliminating occupations, a course can keep its title while the competence behind it changes, and drift accumulates between reviews.
• Four theories yield five dimensions. Human capital, sociotechnical systems, institutional theory with path dependence, and adult learning each explain part of the misalignment; their synthesis yields the five dimensions.
• Six tests replace "whose course stays?" Disciplinary relevance, workforce currency, degree-level rigor, vertical differentiation, stakeholder validation, and AI and workplace relevance turn deliberation into evidence, with five outcomes: retain, revise, integrate, replace, or retire.
• Aggregation is not integration. Substituting courses from adjacent disciplines can widen access while leaving program competencies and workforce outcomes only partly demonstrated.

The legacy curriculum problem

Wang sets the present beside two earlier American inflections, Sputnik in 1957 and the 1983 report A Nation at Risk, without treating them as equivalent. Their shared lesson is structural: change exposes the widening distance between inherited structures and emerging workforce requirements. AI may differ in continuous improvement and reach, which shorten the interval in which an inherited structure stays defensible.

A legacy curriculum is one whose survival is driven by historical arrangements, accumulated practice, established faculty assignments, or administrative convenience rather than demonstrated alignment with contemporary occupational competencies, student needs, and workforce expectations. It is not simply an old curriculum: longstanding courses stay sound while outcomes match the field. The problem begins when continuity itself becomes the justification for continuation.

The manifestations are familiar. Courses persist because they were always offered. Existing faculty expertise decides what is taught instead of contemporary competencies deciding what expertise is needed. Courses borrowed from adjacent disciplines substitute for coursework without evidence of equivalent outcomes. Bachelor's and master's versions carry similar content without vertical differentiation. Individually these look like housekeeping; together they leave a program administratively operational and educationally stale.

The urgency rests on labor-market evidence. An ILO analysis of nearly 30,000 occupational tasks finds that approximately one-quarter of global employment has some exposure to generative AI, with job transformation rather than wholesale replacement the likely consequence. The World Economic Forum is cited for employers expecting 39% of workers' existing skill sets to be transformed or become outdated by 2030. Hence the sharp line: encouraging students to use generative AI, teaching prompting, or writing acceptable-use policy leaves outcomes untouched. That is adoption, not curriculum transformation, because instructors adopt tools quickly while changing outcomes, sequences, assessments, and governance requires coordinated action.

Four theories, five dimensions

The framework comes from theory integration, not model testing. Human capital theory ties education to the productive value of evolving competencies, supplying an external relevance test. Sociotechnical systems theory rejects separating technology from the people and organization doing the work, so curriculum must represent the reorganized system of people, AI, data, and judgment. Institutional theory and path dependence explain persistence: courses occupy permanent slots in degree structures, assignments attach to content areas, and changing it costs more than leaving it. Adult and workforce learning covers learners who arrive with occupational knowledge from employment and should not retraverse what they have mastered.

The synthesis produces five interacting dimensions. Occupational transformation is the external starting point: what has shifted inside occupations through AI, automation, and new human-machine relationships, read through occupational standards. Curriculum currency asks whether outcomes, courses, and assessments still represent those competencies, which requires translation into authentic assessment rather than refreshed terminology. Faculty capability alignment asks whether collective expertise matches what the program promises, separating capability from ownership: curriculum should determine the expertise required, not available expertise determine the curriculum. Program architecture concerns coherence across courses: progression from foundational to advanced competence, vertical differentiation between degree levels, substitution integrity, assessment coherence. External alignment closes the loop through employers, professional bodies, graduates, workforce data, accreditation, and policy.

AI is an accelerator acting across all five, not a sixth dimension. The dynamic runs from occupational change to curriculum renewal, faculty capacity, program coherence, workforce outcomes, external feedback, and back to renewed occupational analysis. Approval of a revised program begins another evidence cycle rather than ending one. Hence the central proposition: the greater the rate of occupational transformation, the greater the institutional capacity required for continuous renewal.

Three states, six tests, five decisions

The three-state model separates continuities that look identical in a catalog. Legacy curriculum persists through history, routines, and staffing, with relevance assumed. Aggregated curriculum assembles or substitutes existing courses for flexibility, but their relation to program competencies, degree-level rigor, and workforce outcomes is only partly demonstrated; it widens access while producing fragmentation. Integrated adaptive curriculum selects courses from defined outcomes, contemporary competencies, stakeholder evidence, and degree-level expectations, and shows integration through competency mapping, coherent sequencing, valid assessment, and recurring feedback. For interdisciplinary programs serving working adults, flexibility is fine if bounded by explicit outcomes and professional relevance. The governing question is what competency a course develops, at what degree level, and on what evidence it remains relevant.

Six tests operationalize the judgment. Disciplinary relevance: does the course belong within the program's contemporary intellectual and professional domain, since historical inclusion is not justification? Workforce currency: does it address competencies required or demonstrably emerging, checked against occupational standards, labor-market data, and employer evidence? Degree-level rigor: does intellectual demand match the credential, with course numbering ruled out as proof? Vertical differentiation: is progression across bachelor's and master's levels demonstrable rather than duplicative? Stakeholder validation: can inclusion be justified through graduates, employers, professional bodies, and accreditation, without external constituencies dictating curriculum? AI and workplace relevance: does it prepare students for human-AI work, meaning AI-mediated tasks, judgment, accountability, and workflows, rather than merely teaching about AI or permitting tool use?

Each component then gets one revisable decision: retain, revise, integrate, replace, or retire. Wang notes that most vocational systems already review every three to five years with supplementary updates, and argues that AI acceleration should strengthen the supplementary function. The posture is stewardship on demonstrated relevance, neither preservation nor disruption for its own sake.

From periodic review to continuous curriculum intelligence

Operationalization starts from a recurring evidence cycle, not a one-time redesign. At an interval matched to the pace of occupational change, leaders assemble evidence per dimension, then apply the six tests course by course to produce a documented decision. Indicators should be observable: the proportion of competencies validated against current occupational standards, the percentage of courses with recently mapped outcomes and authentic assessments, evidence of progression between degree levels, faculty development aligned to capability gaps, and documented closure of earlier gaps. Thresholds are calibrated locally.

A worked hypothetical makes the logic concrete. In an information technology program where troubleshooting is increasingly AI-assisted while demand grows for cybersecurity judgment, data interpretation, AI oversight, and client communication, the framework might retain several courses with revised outcomes, integrate two overlapping ones, target faculty development at AI-enabled workflows, and send a revised competency map to an advisory board. The move is not an AI unit in every course but redesign around the changed configuration of competence. Constraints are acknowledged: faculty time, reliable evidence, governance capacity, and for smaller institutions, phased priorities. Labor-market signals can be volatile, and employer preference should not override durable disciplinary knowledge. The paradigm shift runs from periodic review to continuous curriculum intelligence, supported by AI under human-centered governance and human accountability.

What this means for practice

• Instructors. Adding a generative AI assignment is not renewal. Ask whether your outcomes and assessments still describe the occupation students are entering, and whether the version you teach survives contact with how the work is now divided between people and systems.
• Curriculum designers. Map every course to a named program competency and a degree level before debating content. Aggregation is the easy default when substitutions are approved one course at a time, and it is invisible in a catalog.
• Program leaders and administrators. Replace "can this course count?" with the competency, degree-level, and evidence question, and record an explicit retain, revise, integrate, replace, or retire decision per course, so the reasoning survives staff turnover.
• Faculty developers. Let identified capability gaps drive hiring and professional development rather than existing course ownership, which is how staffing quietly becomes curriculum policy.
• Advisory boards and employer partners. Supply evidence for workforce currency and expect to see how it changed a decision. Employer input is one input, not a veto over disciplinary rigor.

Limitations

• The framework is untested. Wang states that its dimensions and tests have not been empirically validated across institutions, occupational sectors, or governance contexts, and that the relative weight of each dimension may vary by program.
• No evidence of effect. There are no cases, comparison sites, or outcome data, so nothing shows that continuous curriculum intelligence improves currency or workforce alignment, or that it avoids the administrative burden and short-termism the author flags.
• Key constructs remain operational only in principle. Currency, evidential sufficiency, and the weighting of conflicting stakeholder claims are left to local calibration, so two programs applying the same tests could reach opposite decisions and both be defensible.
• The legacy definition risks circularity. Relevance is meant to be demonstrated, but the standards for demonstration are what the framework asks institutions to build, so judgment may fall back on the evidence a program already has.
• Capacity is assumed. The model needs faculty time, reliable labor-market data, and coordination, and the article concedes that smaller institutions may manage only phased cycles; the recommended indicators are process and mapping measures, not learning outcomes.
• Employer evidence is restrained only normatively. The article asserts that labor-market signals should not override disciplinary knowledge and educational purpose without saying who decides when they conflict.

Citation

Wang, V. (2026). When Legacy Curriculum Meets Artificial Intelligence: A Theoretical Framework for Institutional Renewal in Career and Technical Education. International Journal of AI in Pedagogy, Innovation, and Learning Futures, 1(4).