Future-Proof Careers: What Won't Be Obsolete in 10 Years
In January 2025, the World Economic Forum published its Future of Jobs Report: 170 million new positions will be created by 2030 and 92 million will disappear — simultaneously. The question isn't "will my career survive?" It's which tasks inside it will remain human.
The OECD (2023) puts around 27% of jobs across OECD countries at high automation risk. Routine and codifiable tasks go first — data entry, standard administrative work, simple calculations. More durable are roles where non-predictable situations, direct people-management, or physical work in variable environments dominate.
For twenty years the comfortable answer was: pick something involving computers or something creative, and you'll be fine. That answer has worn out. The WEF's 2025 Future of Jobs Report places graphic designers among the fastest-shrinking occupations — alongside cashiers, bank clerks, and data-entry operators. Generative AI can produce standard visual work cheaply and quickly. "Creative" as a category is no longer a safe label on its own.
Where the 47 percent figure came from — and why it's outdated
In 2013, Oxford economists Carl Benedikt Frey and Michael Osborne analyzed 702 occupations from the U.S. Department of Labor database to estimate the probability of computerization for each. Their finding was blunt: roughly 47 percent of U.S. employment is at high risk. The number traveled fast.
The problem is that the study rated entire job titles, not the discrete tasks that make up those jobs. The OECD took a different cut in 2023: what fraction of the actual operations in a given role are automatable? Result: around 27 percent of jobs across OECD countries face genuinely high risk. The difference matters — an accountant who negotiates with clients on top of running figures is in a very different position from one who only runs figures.
What's declining and what's growing — the WEF 2025 picture
The fastest-shrinking roles in the WEF report: postal-service clerks, bank tellers, cashiers, data-entry operators. The common thread is tasks with predictable inputs and outputs — an algorithm can handle those more cheaply than a person can.
The fastest-growing direction in absolute numbers is unexpected: agricultural and green-economy workers, projected to add 34 million new positions by 2030. Also growing: big-data specialists, AI engineers, software developers, nursing and social-care professionals.
Graphic designers are in the shrinking column. That's the gap in the conventional picture.
David Autor showed in the Journal of Economic Perspectives (2015) that automation doesn't just displace workers — it polarizes the labor market. The middle layer — clerks, operators, routine technical workers — shrinks fastest.
The tasks machines still handle poorly
Autor's framework divides work into routine tasks — those that follow a clear algorithm and can be automated — and non-routine tasks, which come in two kinds.
Non-routine cognitive: judgment in unique situations, analysis under incomplete information, solving problems without precedent. Think of the strategic consultant, the surgeon with an atypical case, the teacher managing a difficult class.
Non-routine manual: physical work in unpredictable, changing environments — the plumber, the electrician, the bedside nurse. The OECD (2023) singles out social and interpersonal skills — empathy, managing trust, reading a conversation that's going sideways — as the primary bottleneck to automation. An algorithm cannot talk a frightened patient into making a decision, or notice in real time when the dynamic in the room has shifted.
What this means when choosing a path
The practical takeaway: look at what tasks dominate a role, not just what the role is called. "Lawyer" is too broad — standard contracts are already generated automatically; negotiating with a difficult counterparty is not. "Software developer" is also too broad — boilerplate code is written by AI; architectural decisions in a novel project are not.
A good first step is to understand what type of work pulls you and where you're strongest: taking things apart to understand them, working closely with people, finding solutions where there's no established playbook. That's what career orientation explores — before you commit to a specific program or institution.
What to do right now
- Look at a profession from the inside: what does someone in that role actually do each day? How much of it is routine?
- Find the points where judgment, adaptation, or direct human interaction are required — that's the durable part of any role.
- Take the free career assessment: it will show you which types of tasks you're drawn to and where to go from there.
- Don't choose a career by its name. Choose it by the tasks inside it.
Common questions
Which careers are least likely to disappear in the next 10 years?
According to the WEF Future of Jobs Report 2025, the most resilient roles include healthcare and social-care workers (nurses, psychologists, social workers), AI and big-data specialists, and green-economy workers. The shared trait is tasks that require adapting to unpredictable situations and working directly with people.
Are creative careers protected from automation?
Not automatically. The WEF 2025 report lists graphic designers among the fastest-shrinking occupations — alongside cashiers and data-entry clerks. Generative AI can reproduce standard visual work cheaply and quickly. What holds up is not "creativity in general" but original judgment in genuinely novel situations.
How should I factor automation into a career choice?
Economist David Autor (Journal of Economic Perspectives, 2015) showed that automation primarily targets routine, codifiable tasks. Non-routine work — physical tasks in unpredictable environments and cognitive judgment in unique situations — is more durable. When choosing a path, look at what tasks dominate a role, not just what the job is called.
Sources
- World Economic Forum (2025). Future of Jobs Report 2025. Davos: WEF. weforum.org
- Frey, C. B., & Osborne, M. A. (2013). The Future of Employment: How Susceptible Are Jobs to Computerisation? Working Paper. Oxford Martin School, University of Oxford.
- Autor, D. H. (2015). Why Are There Still So Many Jobs? The History and Future of Workplace Automation. Journal of Economic Perspectives, 29(3), 3–30.
- OECD (2023). OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market. Paris: OECD Publishing. oecd.org
This material is for educational purposes and does not constitute professional advice. Labour market projections are data-based estimates, not guarantees.
Find out which types of work suit you best
15 minutes, no sign-up required. You'll get an interest profile and careers where those tasks are most common.
Take the free test →