Should Every Degree Become an AI-Enabled Degree?

For decades, higher education has organized degrees around the knowledge and professional practices of disciplines: accounting, architecture, biology, business, education, engineering, history, law, medicine and hundreds more. Artificial intelligence now challenges a basic assumption behind that structure: that the ways graduates will research, analyze, communicate, design and make decisions will remain sufficiently stable for curricula to be updated in occasional cycles.

AI is not a separate industry that students can simply choose to enter or avoid. It is a general-purpose family of technologies crossing industry boundaries, though at very different speeds and with very different consequences. A nurse may encounter AI-assisted clinical documentation; an accountant may assess anomaly detection; an architect may evaluate generative design; a historian may challenge machine-produced interpretations of archival material. Their tasks, risks and obligations are not interchangeable.

The central argument is therefore not that every student must learn to code neural networks. It is that every degree should deliberately prepare graduates to understand, evaluate and—where professionally appropriate—work with AI without surrendering disciplinary expertise or human judgment.

The answer calls for an institutional redesign, not a cosmetic change in course catalogs. Adding “Artificial Intelligence” to one module will not adequately prepare graduates if assessments still reward only outputs that tools can produce, faculty lack support, employers are not consulted and professional risks are ignored.

1. The labor market is changing—but exposure is not the same as replacement

The World Economic Forum's Future of Jobs Report 2025 found that 86% of surveyed employers expected AI and information-processing technologies to transform their businesses by 2030. The same report estimated that 39% of workers' core skills would change over the 2025–2030 period. Both figures represent employer expectations from the report's survey—not observed outcomes for all firms or a prediction that 39% of jobs will disappear.[1][2]

Similarly, the International Labour Organization's 2025 task-based assessment found that about one quarter of global employment was in occupations with at least some exposure to generative AI, while 3.3% of global employment fell into its highest exposure category. The ILO emphasizes differences across occupations and income groups and does not equate potential task exposure with the elimination of jobs.[3]

The implication for higher education is precise: universities should plan for changes to tasks within professions, not assume that every profession is equally automated or that any particular tool will dominate future practice.

This geographic variation also matters. An AI-intensive finance hub, a small agricultural community and an economy with limited connectivity cannot rationally adopt identical curricula. Yet the breadth of applications makes basic AI literacy increasingly relevant across settings, provided institutions adjust depth, costs and examples to local circumstances.

2. The pressure is already inside universities

The Higher Education Policy Institute's Student Generative Artificial Intelligence Survey 2026, based on a survey of 1,054 full-time UK undergraduates conducted in December 2025, reported that 95% had used AI in at least one way and 94% had used generative AI to help with assessed work. Only 38% said their institution provided AI tools. These are self-reported UK responses and should not be presented as a global university rate.[4]

There is a widening policy challenge: when students encounter AI before their curriculum addresses it, learning depends on informal experimentation, unequal access and inconsistent rules. Some become sophisticated evaluators; others may rely on polished but unreliable answers. An institution that simply prohibits tools in every task may avoid difficult redesign decisions but may also leave graduates unprepared for legitimate professional use.

In its Digital Education Outlook 2026, the OECD draws an essential distinction: producing a better answer with generative AI does not necessarily mean a student has learned more. Its review notes that unguided AI use can improve immediate output while creating risks of weaker understanding. That makes curriculum and assessment reform inseparable.[7]

3. Does “AI in every degree” mean the same AI course for everyone?

No. A university-wide expectation does not require identical content or identical technical depth. A mathematics major may need model evaluation and statistical assumptions; a law major may need confidentiality, case verification and professional ethics; a design major may need generative workflows, provenance and intellectual property; a teacher may need pedagogically justified AI-supported instruction.

A credible institution-wide policy should therefore separate three decisions: what every graduate must know; what each discipline must be able to do; and which students require specialist AI engineering capabilities. Confusing these levels produces shallow generic courses, unnecessary programming requirements or technology-driven content unrelated to professional tasks.

UNESCO's 2024 AI competency framework for students presents 12 competency blocks across human-centered thinking, ethics, AI techniques and applications, and AI system design, at understand–apply–create levels. It is principally a framework for school education; universities can use its logic as a reference, not as an off-the-shelf higher-education accreditation standard.[8]

4. Start with graduate capabilities, not software licenses

Program restructuring should begin by asking what a competent graduate must demonstrate. Across disciplines, seven outcomes form a defensible starting point. They should be calibrated to the qualification level, jurisdiction, profession and student population.

Proposed synthesis for curriculum design; these are recommended learning outcomes, not a pre-existing formally validated seven-part competency scale.

The OECD's 2025 paper on what students should learn in a future of powerful AI similarly calls for reconsidering not just new competencies, but the relative emphasis, sequencing and learning experiences of existing curricula. This is the difference between adding a topic and redesigning a degree.[9]

5. Restructure the sequence of learning, not merely its content

Curricular change should be visible from the first year through the final major project. Introductory literacy provides a common base, but real competence requires students to encounter AI repeatedly as problems become more complex. A possible four-year progression is shown below. This is an illustrative design proposal, not a universal regulatory requirement.

There are important exceptions. A two-year degree, postgraduate program, apprenticeship, regulated clinical qualification or highly structured licensure pathway may require a different sequence. The principle is vertical integration over time, not imposing a four-year template on every qualification.

6. What should actually change in different degree programs?

The decisive redesign happens inside the major. AI-related outcomes should emerge from job-task analysis, employer dialogue, professional standards, disciplinary scholarship and evidence of real tool use—not the popularity of a product. The table offers curriculum examples for investigation, not claims that every profession already uses these applications at scale.

7. How curriculum committees should decide what to keep, change or remove

Redesign is not synonymous with adding contact hours. Many programs are already crowded, particularly where external accreditation, laboratory requirements or field placements apply. A sustainable review should examine each module at the level of the learning outcome and assessment task.

8. An illustrative 120-credit degree: integrate AI without inflating credits

Consider a hypothetical 120-credit bachelor's program. The institution could deliberately redesign 15 existing credits to carry explicit, assessed AI-related outcomes: six credits in a shared foundation and methods sequence, six across major-specific courses, and three in a culminating project. These are not 15 additional credits. They are a curriculum planning example subject to each program's academic rules and accreditation conditions.

The more important metric is not how many credits mention AI, but whether every graduate encounters appropriate opportunities to learn, practice, apply and demonstrate responsible AI capability. For example, a law program that adds an elective but never requires students to verify citations in an AI-assisted legal memorandum has not necessarily developed professional competence.

9. Redesign assessment before AI makes existing assessments less meaningful

Universities must distinguish two kinds of competence: what students can do independently and what they can accomplish with permitted tools. Both matter. A student who writes an impressive AI-assisted report but cannot explain its reasoning may not have met a disciplinary learning outcome. Equally, banning AI from every activity could prevent assessment of the tools and judgment graduates will actually need at work.

A defensible program-level approach uses three assessment conditions, made explicit in assignment briefs:

The OECD's Digital Education Outlook 2026 supports a selective, pedagogically grounded approach that preserves cognitive effort. This should be reflected in assessment rubrics: reward sound reasoning, source quality, verification, risk management, originality and defensible judgment rather than polish alone.[7]

10. Faculty development is the real capacity constraint

Curriculum reform cannot be delegated entirely to a central AI office or a small group of computer scientists. Faculty members understand disciplinary threshold concepts, professional standards and authentic performance. They also need sufficient preparation to judge which tools support learning and which undermine it.

UNESCO's 2024 AI Competency Framework for Teachers identifies 15 competencies across five dimensions, including AI pedagogy, ethics and professional learning. While it is not a university faculty certification scheme, its structure offers a useful basis for a locally adapted staff development strategy.[12]

Institutions should fund co-design time for discipline faculty, instructional designers, librarians, assessment specialists, IT security, disability support and relevant external practitioners. Effective faculty development involves practice, peer review and teaching redesign—not only demonstrations of emerging software.

11. Governance, fairness and academic quality cannot be optional

Institutional adoption requires rules about permitted uses, access, intellectual property, personal data, academic honesty, evaluation, vendor procurement and security. It also requires a practical escalation route when a model produces harmful or inaccurate outputs. UNESCO's guidance on generative AI emphasizes human-centered policies and data privacy, while NIST's AI Risk Management Framework organizes practice around the functions of govern, map, measure and manage.[13][14]

The OECD's July 2026 policy spotlight on responsible, systematic generative-AI adoption in higher education identifies common areas of government and system response: guidance, shared compliance and procurement approaches, competencies, pilots and evidence gathering, and specialized educational tools.[15]

Affordability is not a peripheral concern. If “AI readiness” effectively means that students must privately subscribe to powerful tools to achieve equivalent grades, universities may deepen inequality. Procurement and course design should protect educational access rather than equate purchasing power with academic performance.

12. A realistic 18-month implementation pathway

Curricular transformation requires clear academic ownership and manageable stages. The following schedule is a planning illustration, not a proven universal implementation timeline. Programs with external accreditation may require longer approval cycles.

13. Measure learning and graduate readiness—not the number of AI workshops

A university could run dozens of AI training sessions and still fail to improve graduate capability. Meaningful evaluation requires a baseline, clear definitions and evidence across both student learning and program quality.

Institutional leaders should report both successes and unintended consequences. If students become quicker at generating content but weaker at reasoning without tools, that is a warning sign, not an educational improvement. Where feasible, compare pilots with credible alternative teaching arrangements rather than treating all improvement as caused by AI integration.

14. Questions university leaders should ask now

15. The strategic choice: AI-enabled education or AI-shaped credentials?

There is a real danger that universities respond to technological pressure with symbolic change: a newly branded course, a collection of optional workshops, a policy that focuses exclusively on cheating, or a promise that every graduate will be an “AI expert.” None of these is a substitute for thoughtful academic design.

A more credible ambition is to ensure that all graduates understand AI's possibilities and limitations; that each major incorporates the applications, verification standards and professional judgments relevant to its field; and that faculty can preserve the foundations of genuine learning as tools evolve. This may mean integrating AI broadly into some degrees and more cautiously into others, including areas where using AI is unsafe, legally restricted or pedagogically counterproductive.

The OECD's 2026 analysis of higher education policy and UNESCO's competency work point toward responsible, evidence-led institutional responses rather than tool-led enthusiasm. The appropriate question for university leaders is therefore no longer whether AI deserves a place somewhere in the curriculum. It is how every degree can develop the right level of AI capability without weakening the knowledge, ethics and judgment that justify a university education in the first place.[15][8]