Short answer: students should prepare for careers in the AI era by pairing a real field of knowledge, whether that's accounting, nursing, design, logistics or law, with confident, honest use of AI tools, and by building skills that AI doesn't replace: clear thinking and writing, judgment, working with people, and learning quickly. Then prove it with a small portfolio of real projects. You don't need to become an AI engineer unless that's what genuinely interests you.
It's a stressful time to be choosing a path. Headlines swing between "AI will take every job" and "learn to prompt and you're set," and neither is useful. This post is a calm, practical guide: what employers are likely to look for, which skills last, how to use AI in your studies without shortchanging yourself, and how to get experience when the traditional entry-level tasks are shrinking.
What will employers actually look for?
Employers will hire people who can produce good work with AI, not people who avoid it and not people who hand everything to it. From the employer's side, that looks like:
- Real knowledge of the field. Someone who understands the work well enough to tell when AI output is wrong.
- Fluency with tools. Comfort using AI assistants and AI features in everyday software, without needing hand-holding.
- Checking habits. Evidence that you verify facts and figures rather than trusting confident-sounding answers.
- Communication. Writing and speaking clearly, especially explaining complex things simply.
- Initiative. Spotting a slow process and improving it, which is exactly what businesses need as they adopt AI.
That last point is worth stressing. Many small and mid-sized businesses are still figuring out how to use AI. A graduate who arrives able to say "I can help you automate this report" is valuable in a way that's hard to fake. For a sense of where businesses are applying AI, browse AI use cases by department.
Which skills last no matter how AI changes?
Tools will change many times over a career. These skills transfer across all of them:
- Clear thinking and writing. If you can define a problem precisely, you can direct AI precisely. Muddled thinking produces muddled prompts and muddled results.
- Number sense. Not advanced math necessarily, but knowing when a figure looks wrong, and how to check it.
- Judgment and verification. AI can be confidently wrong. Our explainer on why AI makes things up shows what that looks like in practice. Being the person who catches it is a career skill.
- Working with people. Persuading, negotiating, teaching, caring, leading. These are the parts of most jobs AI touches least.
- Learning how to learn. The ability to pick up a new field, tool or method quickly may be the single most durable skill of all.
- Ethics and responsibility. Knowing when something shouldn't be automated, and speaking up about it.
None of these are new. What's changed is that they're now the main thing that distinguishes you, because AI handles more of the mechanical parts of work.
Why does "domain plus AI" beat "just AI"?
"Good with AI" is becoming as common as "good with computers" was a generation ago. On its own, it won't set you apart for long.
What's rare is the combination: a nurse who can design a better shift-handover process with AI, an accountant who can automate reconciliations safely, a logistics graduate who can get AI to read shipping documents, a marketer who knows the brand well enough to edit AI drafts into something good. In each case, the field knowledge is what makes the AI useful.
So choose a field you find genuinely interesting, learn it deeply, and treat AI as a power tool within it. That's a more robust plan than chasing whichever "AI job" is fashionable this year. For a framework on which kinds of work are most and least exposed, see which jobs are safest from AI.
Should you study AI or computer science?
Only if it genuinely interests you. Computer science, data science and engineering remain valuable, and people who build AI systems will be in demand. But you don't need a technical degree to have an AI-era career, and choosing one purely out of fear tends to go badly.
If you're in a non-technical field, a little technical literacy still helps: understanding what data is, how spreadsheets and databases work, what an automation does, and how to read a basic chart critically. Many students now build small tools by describing them to AI, sometimes called "vibe coding." That's a great way to learn, as long as you understand its limits, which we cover in vibe coding for business.
How do you use AI in your studies without cheating yourself?
First, follow your institution's rules. They vary widely, from course to course and country to country, and breaking them can have serious consequences.
Beyond the rules, there's a practical reason to be careful: if AI does your thinking, you don't learn, and you'll be competing later against people who did. Useful ways to use AI while studying:
- As a tutor: ask it to explain a concept another way, or to quiz you.
- As a sparring partner: write your argument first, then ask AI for the strongest counter-arguments.
- As an editor: get feedback on clarity after you've written the draft yourself.
- As a research starter: get an overview, then check every claim against proper sources.
Ways that hurt you: having it write assignments you submit as your own, or accepting answers you haven't checked. The struggle of working something out is often where the learning actually happens.
What should go in an AI-era portfolio?
A portfolio is a small collection of real work that shows what you can do. In the AI era, the most convincing projects show a real problem, how you used AI, how you checked it, and what improved.
| Field | Project idea | What it shows |
|---|---|---|
| Business or commerce | Automate a weekly sales report for a local shop or society | Process thinking, data handling, initiative |
| Marketing | Run a content plan for a real club or small business with AI drafts you edited | Brand judgment, editing, results tracking |
| Accounting or finance | Build a checked AI workflow that categorizes expenses | Accuracy, verification, domain knowledge |
| Healthcare or social care | Design patient-friendly explanations of a procedure, reviewed by a professional | Communication, responsibility |
| Engineering or computing | Build a small tool that solves a real annoyance, with tests | Building, reliability |
| Humanities or law | A research brief where you document every AI claim you verified or rejected | Critical thinking, rigor |
Write a short note for each: the problem, what you did, what AI did, what you checked, what went wrong and what you'd do differently. That honesty is more impressive than polish.
How do you get experience when entry-level tasks are automated?
The routine tasks juniors traditionally learned on are shrinking, so you have to be more deliberate:
- Help a small business or nonprofit. Many need AI help and can't afford consultants. A semester project for a real organization is excellent experience.
- Look for internships that involve people. Customer contact, sales support, operations. These build judgment AI can't give you.
- Freelance small, real jobs. Even modest paid work teaches you about clients, deadlines and accountability.
- Join or start a project. Student societies, competitions, open-source projects and community initiatives all count.
- Ask for feedback from practitioners. People working in your field can tell you what's changing faster than any article.
If you're worried about specific roles, our employee-focused guide, will AI take my job?, includes a self-check you can apply to careers you're considering.
Frequently Asked Questions
Is it still worth going to university in an AI world?
For many people, yes. A degree gives you deep knowledge of a field, practice at learning hard things, and networks that matter. What's changing is that the degree alone is less of a signal; employers increasingly want to see what you can actually do, which is where projects and experience come in.
Which degree is safest from AI?
No degree is AI-proof, and choosing by fear rarely works out. Fields involving physical presence, care, trust and accountability are generally less exposed. But a strong student in an "exposed" field who uses AI well can do better than a weak student in a "safe" one.
Should every student learn to code?
Not necessarily, but basic technical literacy helps everyone: understanding data, spreadsheets and how automation works. If you enjoy coding, it remains a valuable skill. If you don't, focus on your field and learn enough to work alongside technical people.
Will employers mind that I used AI on my projects?
Most will care more about whether the work is good and whether you understand it. Be open about how you used AI and how you checked the results. Hiding it, then being unable to explain your own work in an interview, is the real risk.
How do I keep up with AI changes while studying?
Use AI regularly on real tasks in your field, and follow a few practitioners rather than a flood of news. Tools will change; the habits of trying, checking and reflecting carry over.
Running a business that wants to put AI to work?
Students and graduates often become the people who bring AI into the businesses they join. If you run a business and want help working out where AI fits, tell us about your team. We help small and mid-sized businesses find practical first uses and automate them safely. See how we approach AI automation.