A degree can still be worth the time and money in the age of AI, but that answer becomes unreliable when applied to every student, course, country, or career.
Current Organisation for Economic Co-operation and Development (OECD) data still show a substantial average employment and earnings advantage for tertiary education. At the same time, recent evidence points to a more difficult early-career market in some countries and possible pressure on young workers in occupations with high exposure to AI. Those findings matter, but they do not establish that AI has made degrees obsolete.
For someone deciding whether to enrol, continue studying, change direction, or pay for another qualification, the useful decision is more specific: Does this particular degree provide access to work you want, build capabilities employers or professional bodies value, produce credible work evidence, and justify its full cost and completion risk?
That is a harder question than “college or no college,” but it is also far more useful.
Answer Summary: Degrees still have substantial average labour-market value, but a degree is not automatically a good investment. AI is changing tasks and may be affecting some early-career hiring, yet exposure to AI does not equal job loss. A sound decision checks five areas: whether the credential is required, what local outcomes show, the full cost and likelihood of completion, the practical evidence you will graduate with, and whether the learning remains useful as tools change.
Table of Content
- What does current evidence say about degree value?
- What has AI changed about the decision?
- Have employers stopped caring about degrees?
- When is a degree difficult to replace?
- When do certificates, apprenticeships, and portfolios deserve comparison?
- Why completion risk changes the value of a degree
- Five checks before spending years and money on a degree
- How does the answer change for different readers?
- Mistakes that distort the degree decision
- Conclusion
Key Takeaways:
- Tertiary education still has a strong average employment and earnings advantage across OECD countries.
- Those averages describe groups, not the return from one program for one person.
- AI exposure means tasks may change; it does not show that an occupation will disappear.
- Recent U.S. evidence raises concerns about young workers in some highly exposed occupations, but causation remains unsettled.
- Removing degree requirements from job advertisements does not necessarily change who employers hire.
- Completion, recognition, work experience, total cost, and local demand can change the value of the same type of degree.
What does current evidence say about degree value?
The broad evidence still favours tertiary education on average, but the averages hide large differences between countries, qualifications, fields, and individual circumstances.
Employment remains higher among tertiary-educated adults
In the OECD's 2026 data, 87% of tertiary-educated adults aged 25–64 were employed. The corresponding employment rate was 78% for adults with upper-secondary or post-secondary non-tertiary education and 60% for those below upper-secondary attainment. Among tertiary-educated 25–34-year-olds, employment was also 87% in 2025.
That is strong evidence against the claim that degrees have lost all labour-market value.
It is not evidence that an individual student has an 87% chance of getting a suitable job after completing any degree. The OECD figures combine many qualifications and national labour markets. A program in a weak local market, an unrecognised qualification, or a course with poor completion outcomes can produce a different result.
The earnings advantage remains sizeable, but it is not a personal forecast
Among full-time, full-year workers across OECD countries, the 2026 report found average earnings advantages over workers with upper-secondary education of 18% for short-cycle tertiary qualifications, 39% for bachelor's degrees, and 80% for master's, doctoral, or equivalent qualifications. Tertiary-educated workers aged 25–34 earned 38% more on average than upper-secondary peers; for bachelor's holders in that age group, the difference was 31%.
Country differences are substantial. For tertiary-educated 25–34-year-olds, the OECD reports much smaller premiums in some countries and far larger ones in others.
The wording matters here. These statistics show an association between educational attainment and earnings. They do not prove that earning one specific degree causes an individual person's income to rise by the OECD average. Field of study, previous education, labour demand, work experience, geography, and other factors also shape outcomes.
A prospective student asking, “Will this degree pay?” needs local and program-specific evidence, not an international average presented as a promise.
Skills still separate outcomes among people with similar credentials
A qualification and capability are related, but they are not identical signals.
OECD analysis using the Survey of Adult Skills found that about 65% of adults with high proficiency across literacy, numeracy, and adaptive problem-solving were in the top two earnings quintiles. Among tertiary-educated adults across all measured skill levels, the equivalent figure was 54%. The OECD concludes that both educational attainment and measured skills are positively associated with labour-market success.
This is an important distinction for the AI-era degree debate. A degree can retain value as a qualification, screening signal, and route into certain occupations while still being insufficient evidence that someone can perform well at work.
The stronger position is not “degrees beat skills” or “skills beat degrees.” The two signals do different jobs.
What has AI changed about the decision?
AI has made task change and early-career hiring more important parts of the degree decision. It has not produced evidence for a universal conclusion that university education no longer matters.
AI exposure does not mean a job will disappear
The International Labour Organization's 2025 global assessment estimated that one in four workers worldwide was employed in an occupation with some degree of exposure to AI systems capable of generating content. The ILO's interpretation is equally important: because human input remains necessary across most affected occupations, transformation of jobs is more likely than wholesale replacement.
“Exposure” is often misread.
An occupation can be exposed because some of its tasks can be performed or assisted by AI. That does not show that employers will automate those tasks, eliminate the occupation, reduce headcount, or stop hiring graduates.
A student assessing a field needs to ask what is changing inside the job. Which tasks are becoming easier to automate? Which require judgment, accountability, physical work, client interaction, regulation, or contextual knowledge? How are employers reorganising the work around new tools?
Those questions are more informative than asking whether an AI system can produce a piece of text, code, analysis, or design resembling one part of the occupation.
The early-career warning is real, but its cause is not settled
There is evidence of recent pressure on young tertiary graduates.
Across the OECD, unemployment among tertiary-educated 25–34-year-olds rose by 0.4 percentage points between 2023 and 2025, while unemployment among young adults with upper-secondary attainment fell by 0.2 points. The unemployment gap between the groups narrowed from 2.4 to 1.8 percentage points. The OECD explicitly cautions that evidence for a lasting AI-driven change in skill demand remains limited.
A 2026 Stanford Digital Economy Lab analysis gives a narrower U.S. signal. Using ADP payroll data through June 2026, researchers found no evidence of widespread economy-wide job displacement. They did find that employment among workers aged 22–25 in highly AI-exposed occupations was about 19% below the level it would have reached if it had kept pace with employment among similarly aged workers in less-exposed occupations. The divergence appeared mainly through weaker hiring rather than increased separations.
That 19% figure should not be read as “AI eliminated 19% of young workers' jobs.” It is a relative employment gap in one U.S. payroll dataset. The researchers describe the evidence as descriptive and discuss alternative explanations and measurement limitations.
The Federal Reserve Bank of New York provides another U.S.-specific signal. In the second quarter of 2026, unemployment among recent U.S. college graduates was about 5.6%, while underemployment was about 42%. The New York Fed defines underemployment as a college graduate working in an occupation that typically does not require a college degree. Its “recent graduate” group covers ages 22–27.
These numbers describe a difficult U.S. entry market. They do not identify AI as the sole cause, and they should not be applied to graduates in other countries without local evidence.
AI can change productivity without removing the worker
Technology can substitute for some work while assisting people with other work.
A 2025 study in the Quarterly Journal of Economics examined 5,172 customer-support agents using an AI assistant. Access to the tool increased issues resolved per hour by 15% on average, with larger gains among less experienced and lower-performing workers. The study covered one company and one type of work, so its results should not be applied to every occupation.
For education decisions, the useful lesson is narrower: the effect of AI depends partly on how a job combines human knowledge with the technology, not only on whether software can perform isolated tasks.
Have employers stopped caring about degrees?
Some employers have removed degree requirements, but evidence from large U.S. firms shows that changing job advertisements and changing hiring behaviour are not the same thing.
A 2024 Burning Glass Institute and Harvard Business School study examined 11,300 roles at large U.S. firms using job-posting data and career histories covering 65 million workers. In roles where employers removed degree requirements, the share of workers hired without a bachelor's degree increased by about 3.5 percentage points on average. Yet only 3.6% of the roles in the analysis had dropped degree requirements during the relevant period.
The same study found that about 45% of firms that removed requirements showed no meaningful difference in the degree mix of their hires.
The implication is practical. A posting that says “degree not required” tells you that one formal barrier has been removed. It does not tell you whether applicants without degrees receive interviews at the same rate, pass employer assessments, or get hired at similar rates.
Readers who need a broader treatment of this issue can see Collegenp's coverage of skills-first hiring and degree versus experience.
When is a degree difficult to replace?
A degree is hardest to substitute when it serves as a formal gate rather than only a source of knowledge.
The exact gate varies by country and occupation. Some professional, licensing, registration, or postgraduate pathways require a recognised degree or an approved course of study. In those situations, knowing the material through self-study does not provide the same legal or institutional access.
A student considering such a path should check the responsible regulator, licensing authority, professional body, or admissions office in the country where they intend to work or continue studying.
The key checks are specific:
- Does the occupation require a particular qualification?
- Must the program itself be recognised or accredited?
- Is supervised practice required?
- Are additional examinations or registration steps involved?
- Will the qualification be accepted in the jurisdiction where you expect to work?
A broad article cannot answer those questions for every profession and country. They need official local verification before enrolment.
When do certificates, apprenticeships, and portfolios deserve comparison?
Alternative routes deserve serious comparison when the target work does not require a degree and employers have credible ways to judge capability through other evidence.
Terminology differs across education systems. A degree normally represents a structured tertiary qualification with substantial breadth or depth. A certificate often covers a narrower area of learning, though its level and recognition vary. Professional certification may refer to a separate assessment or industry credential rather than an academic program.
Apprenticeships and structured workplace training can combine learning with employment. Portfolios and work samples can provide direct evidence in fields where employers are able to inspect a candidate's output.
The comparison should focus on what each route permits, not its label.
A short credential that costs less and takes less time is not automatically better if it cannot satisfy the target employer's screening rules or professional requirements. A degree is not automatically better because it lasts longer.
Useful comparison points include:
- recognition by relevant employers or authorities;
- depth and relevance of the curriculum;
- assessment quality;
- access to practical or supervised work;
- progression to further study;
- evidence of graduate or participant outcomes;
- transferability of credits where relevant;
- total time and cost.
For a fuller comparison of these routes, see Collegenp's certificate versus degree guide.
Why completion risk changes the value of a degree
Graduate earnings statistics describe people who attained the qualification. A student deciding whether to enrol also needs to ask what happens if completion takes longer or does not happen.
OECD 2025 data from countries and economies with available cohort information found that 43% of students entering bachelor's programs completed a tertiary qualification within the program's theoretical duration. The figure reached 59% when one additional year was allowed and 70% three years after the theoretical end date. Completion patterns differed substantially between education systems.
This matters because the advertised length of a degree is not necessarily the time every entrant spends studying.
Before enrolling, investigate:
- completion rates and typical time to completion, where reliable data exist;
- academic and financial support available to students;
- whether credits can transfer to another program;
- whether leaving early produces any recognised intermediate qualification;
- whether the program can be continued part-time if circumstances change.
For a student already partway through a degree, the calculation changes. Money and time already spent cannot be recovered. The useful comparison is between the remaining commitment and the realistic alternatives available from now on.
Five checks before spending years and money on a degree
A sound degree decision can be organised around five checks: gate, gain, cost and completion, proof, and flexibility.
| Check | What to verify | Why it matters |
|---|---|---|
| Gate | Credential, licensing, registration, employer, or postgraduate requirements | An alternative route has limited value if it cannot open the required door |
| Gain | Local employment, earnings, progression, and field outcomes | International averages may differ sharply from your market |
| Cost and completion | Total cost, study time, financing, likely completion time, and non-completion risk | The value changes if the program takes longer or interrupts earnings |
| Proof | Internships, placements, research, labs, projects, assessed work, or portfolios | A qualification is stronger when employers can see applied capability |
| Flexibility | Foundational knowledge, domain judgment, communication, evidence evaluation, and current tool use | These capabilities can remain useful when particular tools or tasks change |
1. Check the gate before comparing universities
Start with the work you want to enter.
If the occupation has a formal qualification rule, verify it before comparing tuition, rankings, or course features. If no formal gate exists, examine whether employers treat a degree as required, preferred, or optional.
Those words describe different barriers.
2. Look for outcomes close to your real market
International evidence can show the broad value of education. It cannot tell you whether a particular course in your country leads to the work you want.
Where available, examine official graduate-outcome data, national labour statistics, regulator information, and transparent institution-level results. Treat testimonials as individual experiences, not outcome data.
3. Calculate the commitment, not only the tuition
The relevant cost can include fees, materials, living or commuting expenses, financing costs where applicable, and income you give up by studying instead of working.
Time also matters. A program that takes longer than planned may change the financial comparison substantially.
This is general educational information rather than personal financial advice; borrowing and financing decisions depend on local terms and individual circumstances.
4. Ask what evidence of work you will leave with
A course description tells you what will be taught. It does not automatically tell an employer what you can do.
Depending on the field, useful evidence may come from internships, co-ops, supervised placements, laboratories, research, assessed projects, studios, employer-linked work, or portfolio pieces that students are permitted to share.
When reviewing a program, ask what students have completed by graduation besides examinations and course credits.
5. Look for both foundations and current practice
A course focused narrowly on today's software can age quickly. A course that teaches theory while ignoring how the field currently operates can leave graduates with a different problem.
Look for a combination of domain knowledge, reasoning, communication, evidence evaluation, and practice with relevant tools. Collegenp's discussion of future-of-work skills provides a separate treatment of skill development.
How does the answer change for different readers?
The same degree can represent a different decision depending on where you are starting.
If you have not enrolled yet
You have the widest range of choices, so begin with the target occupation rather than the institution.
Suppose a student is comparing a degree with a shorter training route for a job that accepts work samples and technical assessments. The shorter route deserves investigation because the employment gate may not require a degree.
Change the situation to a profession where the local regulator requires an approved qualification, and the comparison changes immediately. A cheaper course that does not satisfy the professional requirement is not an equivalent substitute.
These are illustrative situations, not documented cases. Their purpose is to show why the occupational gate should be checked first.
If you are already studying
Do not treat an AI system completing some coursework-like tasks as evidence that the entire qualification has lost value.
Examine the remaining tuition and time, the role of the credential in your intended career, transferable credits, practical experience you can still gain, and what realistic alternative route is available.
A student halfway through a program faces a different choice from someone deciding whether to enrol. The relevant costs are the costs still ahead, not a desire to justify what has already been spent.
If you recently graduated and are struggling to find work
A difficult job search does not tell you by itself whether the degree was worthless or whether AI caused the problem.
Your local economy, field, work experience, hiring cycle, employer filters, location, and ability to demonstrate relevant work can all affect the outcome.
Compare your qualifications with current requirements for the roles you are targeting. If employers ask for experience, projects, technical assessments, or particular tools, the immediate problem may be evidence of capability rather than another academic credential.
If you are changing careers
Identify the entry barrier before assuming you need another full degree.
An existing degree plus targeted study, a recognised professional credential, workplace training, or portfolio evidence may satisfy some career changes. Other occupations have qualification rules that make additional formal education necessary.
The shortest route is useful only when it reaches the same destination.
If you are paying for someone else's education
Ask questions that reveal what the program provides beyond admission.
What proportion of entrants finish? What practical experience is built into the course? How are students assessed? Is the qualification recognised where the student intends to work? What evidence exists for graduate outcomes? What happens if the student changes field or leaves before finishing?
Those questions are more informative than relying on institutional reputation alone.
Mistakes that distort the degree decision
Several common comparisons produce misleading answers.
Treating an OECD average as your personal return
OECD data establish that tertiary education remains associated with better average labour outcomes. They do not establish that every program offers the same value.
Treating AI exposure as a forecast of unemployment
Exposure identifies tasks that technology may affect. It does not tell you how quickly firms will adopt the technology, how jobs will be redesigned, or whether employment will rise or fall.
Assuming “degree optional” means graduates have no advantage
Job-advertisement wording and actual hiring can differ. The Harvard and Burning Glass evidence shows why both need attention.
Comparing only tuition
A fair comparison includes time, completion risk, delayed earnings, recognition, and what employment options each route opens.
Searching for a degree that is immune to technological change
No evidence in the research base supports ranking degrees as permanently protected from AI. Occupations contain changing sets of tasks, and exposure differs by technology, employer, country, and job design. The better question is whether the program develops domain knowledge and work evidence that remain useful as those tasks change.
Conclusion
Degrees have not lost their economic relevance because AI can perform more tasks.
Current OECD evidence still shows substantial average employment and earnings advantages for tertiary-educated adults. The ILO's global work shows that AI exposure should not be confused with automatic job elimination. Recent U.S. evidence does, however, give students a reason to take early-career hiring risk and practical work preparation seriously.
The decision becomes clearer when you stop asking whether degrees are worthwhile as a category.
Check whether the credential opens a necessary door. Examine outcomes in the market where you expect to work. Count the full time and cost, including the possibility of delayed completion. Find out what evidence of applied work you will graduate with. Then ask whether the program builds both durable knowledge and the ability to work with changing tools.
For some readers, that will support a degree. For others, it will make a shorter or work-based route worth comparing. The evidence supports making that distinction rather than forcing one answer onto every career.
