7 Entry-Level Jobs AI Is Already Changing, and What to Learn So You're Still Hired

Photo: Tima Miroshnichenko / Pexels

If you're about to start your career, you've probably seen headlines saying AI is wiping out junior jobs. A new study of 41 countries, a working paper published by Stanford's Digital Economy Lab on 21 September 2026, has been reported that way.

The research is worth reading properly, because what it found is more useful than the headline.

What the study found, and what it didn't

The researchers analysed 1.25 billion job adverts and 154 million employment records from 41 countries between 2021 and March 2026. They compared the local offices of multinational companies that had started hiring for generative AI skills with similar offices of companies that hadn't.

At the companies using AI:

• The share of junior staff fell by 1.9 percentage points compared with similar companies.
• That was mainly because senior employment grew, by 6.7%.
• Junior employment was 2.5% lower, but that estimate wasn't statistically significant, so the study can't say for certain that junior jobs fell at all.
• Total employment rose by about 3.3%, though the authors describe this as only marginally significant.

The authors describe AI as saving labour for junior workers and expanding it for seniors in AI-exposed occupations.

So the finding isn't "your job is gone". It's that companies adopting AI are growing by hiring more experienced people, and beginners make up a smaller part of their workforce. That makes the first step into a career harder, not impossible.

One detail worth knowing: South Africa is in the study. It's one of the 31 countries with enough data for its own estimate, and it wasn't among the seven where the decline in the junior share was statistically significant.

Why this matters in South Africa

The first job is where a qualification turns into experience. Writing in Daily Maverick, Prof Linda Meyer notes that nearly six in 10 unemployed young people have no previous work experience. Any squeeze on beginner roles lands hardest here.

The good news: employers aren't only cutting. They're changing what they expect from beginners. That's something you can prepare for.

7 entry-level jobs AI is already changing

Commentators and recent surveys point to these roles as among the most exposed. Three come from Prof Meyer's Daily Maverick column (customer service, data capture and admin, and junior analysts), three from Coface's global study of the fields most exposed to automation (finance, law and creative work) and one from OfferZen's survey of South African developers. For each one, here's what's changing and what to show an employer instead.

1. Customer service and call-centre agents

What's changing: AI chat tools can now answer standard customer questions and draft routine replies.

What to show instead: your ability to handle the cases the tools can't, such as complaints, unusual problems and upset customers. Examples of calm problem-solving count for more than typing speed.

2. Data capture and admin coordinators

What's changing: Prof Meyer lists data capture clerks and administrative coordinators among the first-foothold roles most exposed to automation.

What to show instead: accuracy and judgement. Being able to spot when an automated report is wrong, and fix it, is the skill that remains.

3. Junior bookkeepers and accounts clerks

What's changing: routine capturing, matching and reconciling are increasingly automated in accounting software.

What to show instead: understanding why the numbers look the way they do. Explaining a variance, catching an error and communicating it clearly are what employers still need people for.

4. Junior analysts

What's changing: first-draft research, summaries and basic data clean-up are exactly what generative AI tools are good at.

What to show instead: the questions you'd ask, not just the answers you'd find. A short portfolio piece where you checked an AI summary against the original data and found something it missed is strong evidence.

5. Paralegals

What's changing: first drafts of standard documents and basic research are being done faster with AI tools.

What to show instead: attention to detail and verification. Legal work depends on getting it right, and someone has to check the tools.

6. Junior marketing and content roles

What's changing: first drafts of posts, captions and blog content can now be produced in seconds.

What to show instead: taste and results. Show work where you improved an idea, understood an audience or can point to what happened after it was published.

7. Junior software developers

What's changing: AI coding assistants handle a lot of simple code. In OfferZen's 2026 survey, 48% of tech leaders said they feel pressure to hire more senior engineers as teams get leaner, and 55% said AI fluency and product thinking are now baseline expectations for engineers.

What to show instead: projects where you understood the problem, made design decisions and can explain what you'd change. Being able to spot when an AI tool got it wrong is part of the job now.

Where the human side still protects you

A study by credit insurer Coface, reported by Business Day, found that jobs in care, education, sales and other people-facing work have some tasks at risk too, but their human side acts as a protective factor. If you enjoy working with people, that's worth weighing up when you choose your path.

Three things to put on your CV now

1. Show that you use AI tools well. Say which tools you use and for what, such as drafting, analysis or checking code. HyperionDev's CEO told Business Day that employers increasingly want candidates who can use these tools to be more productive.
2. Show judgement, not just output. Describe a time you checked, corrected or improved something, whether an AI draft or anyone else's work.
3. Show proof. A portfolio, a project, a before-and-after or a result. As HyperionDev's CEO put it to Business Day, employers want proof that you can build, solve problems and adapt.

(See our guide to writing a graduate CV that gets past screening software.)

The bottom line

AI isn't closing the door on young people, but it is moving it. Beginners who can use the tools, check their output and explain their thinking are the ones companies still want to hire. Start building that evidence now, while you're studying or in your first role.