On March 17, Australia's Assistant Minister for Productivity Dr Andrew Leigh stood up at QUT in Brisbane and said something that should worry every uni student in the country. A degree, he argued, might no longer be the career shield it once was.
"The more relevant distinction in the future might be differences in the type of cognition performed," Leigh told the Economic Society of Australia. Not what qualification you hold — but whether your work requires genuine judgment.
It's a striking claim from a former economics professor who holds a PhD himself. But does the data back it up? We looked at 358 Australian occupations to find out which degree-holders face the highest AI risk — and which qualifications still hold their value.
The degree premium is under pressure
For decades, the formula was simple: get a degree, earn more, stay employed. And on paper, it still works. ABS data from 2025 shows 80% of Australians with a post-school qualification are employed, compared with 58% without one.
But dig deeper and the picture gets murkier. For men, full-time employment rates are almost identical whether they hold a bachelor degree (71%) or a certificate III/IV (70%). The gap that once justified three or four years of HECS debt is narrowing — fast.
Employers are noticing too. PwC's Global AI Jobs Barometer found that globally, the share of AI-augmented roles requiring a degree dropped from 66% to 59% between 2019 and 2024. For roles AI can automate, it fell from 53% to 44%. In Australia, LinkedIn data shows a similar slide — from 74% to 69%.
The trend is clear. Degrees are losing their gatekeeping power, and AI is accelerating the shift.
Degree-required jobs with the highest AI risk
Here's where Leigh's argument really hits home. Some of the occupations most exposed to AI are the ones that traditionally demanded a university education.
| Occupation | AI Risk Score | Typical Qualification |
|---|---|---|
| Accountants | 7.3 | Bachelor degree |
| Financial Dealers | 7.1 | Bachelor degree |
| Software Programmers | 6.7 | Bachelor degree |
| Insurance Agents | 6.5 | Bachelor degree / diploma |
| Advertising & Marketing Professionals | 6.2 | Bachelor degree |
| Public Relations Professionals | 5.3 | Bachelor degree |
| University Lecturers | 5.1 | Postgraduate degree |
These are roles where AI can already draft reports, analyse data, write copy, model risk, and generate code. The cognitive tasks that degrees were designed to certify — writing, analysis, pattern recognition — are exactly the tasks large language models do well.
Jobs and Skills Australia's Generative AI Capacity Study confirmed this pattern: the technology is "more likely to enhance workers' efforts in completing tasks, rather than replace them, especially in high-skilled occupations." But the operative word is tasks. If AI handles the tasks your degree trained you for, what's the degree worth?
The trades tell a different story
Compare those numbers with occupations that typically require a TAFE certificate or apprenticeship:
| Occupation | AI Risk Score | Typical Qualification |
|---|---|---|
| Concreters | 1.9 | Certificate III |
| Plumbers | 2.8 | Certificate III |
| Electricians | 2.8 | Certificate III |
| Carpenters & Joiners | 3.1 | Certificate III |
| Motor Mechanics | 3.5 | Certificate III |
AI can write a contract but it can't pour a slab. It can draft a marketing plan but it can't rewire a switchboard. The physical, hands-on, context-dependent nature of trade work creates a barrier that current AI simply cannot cross.
And the demand is there. Most of these trades are rated "Shortage" by Jobs and Skills Australia. Electricians average $154,000 in FIFO roles. By 2026, nine in ten new jobs are expected to need post-school qualifications — and 80% of those are attainable through TAFE.
That's not to say everyone should drop out of uni and pick up a trowel. But it does challenge the assumption that a degree is automatically the safer path.
Not all degrees are equal
The picture isn't uniformly bleak for degree-holders. Healthcare qualifications, in particular, remain strong.
Registered nurses score 4.1 on the AI risk scale, with 362,900 employed and rated "Shortage." Physiotherapists score 4.1 with 19.7% employment growth over five years. Dental practitioners come in at 3.3. Solicitors sit at 4.2, with the profession growing and also rated "Shortage."
What these roles share isn't just a degree requirement — it's hands-on human contact, ethical judgment, and responsibility that can't be delegated to a model. A nurse assessing a patient, a physio adjusting treatment mid-session, a solicitor advising a client through a custody dispute — these involve the kind of cognition Leigh calls "meta-skills."
The dividing line isn't degree versus no degree. It's whether your work requires judgment that carries real consequences.
What Leigh calls "meta-skills"
Leigh's speech identified four capabilities he believes will matter more than credentials:
- Framing problems — knowing which question to ask, not just how to answer it
- Identifying errors — spotting when AI output is wrong, biased, or incomplete
- Allocating attention — deciding what matters when information is abundant and cheap
- Bearing responsibility — being accountable for decisions in ways a model cannot be
These aren't skills you pick up from a textbook. They're built through experience, professional practice, and working in environments where mistakes have real stakes. A graduate accountant reviewing AI-generated tax returns needs to know when the numbers don't smell right. A marketing manager needs to judge whether an AI-drafted campaign will land with a real audience or fall flat.
"Inequality may have less to do with years of education and more to do with whether or not someone occupies a judgment-intensive role," Leigh said.
What this means for your career
The JSA Generative AI Capacity Study found that by 2050, the occupations losing the most jobs will include office clerks, receptionists, bookkeepers, and professionals in sales, marketing, and programming. The largest gains? Cleaners, nurses, construction labourers, and hospitality workers — roles that require physical presence and human connection.
McKinsey estimates 1.3 million Australian workers (about 9% of the workforce) may need to transition into new roles by 2030. The Pearson Lost in Translation report puts it more starkly: 26% of Australian jobs are at high risk if workers don't upskill.
None of this means degrees are dead. Medicine, nursing, law, engineering — fields with professional registration requirements — aren't going anywhere. But for the growing number of graduates in business, marketing, communications, and generalist IT roles, the degree alone is no longer enough.
The PwC AI Jobs Barometer found AI-skilled workers globally earn a 56% wage premium. In Australia, the premium is smaller — around 6% — but it's growing. The workers who thrive won't be the ones with the longest CVs. They'll be the ones who can work alongside AI, spot its blind spots, and make decisions it can't.
Check your own occupation
Whether you're a current student weighing your options, a mid-career professional wondering about your field, or a parent helping a teenager choose a path — the smartest thing you can do is look at the data for your specific occupation.
Use our How Safe Am I? quiz to get a personalised AI risk assessment, or browse the full occupation rankings to see where different roles sit on the scale. You can also compare occupations side by side to weigh up your options.
A degree can still be worth every cent. But in 2026, the question isn't just what did you study — it's what can you do that AI can't?