JENI O’DOWD: If AI takes over all our grunt work, how will next generation of professionals learn?

Mention AI around one of my friends and settle in. You’ll be there for a while.

She thinks it is making people lazy, letting students cheat and sucking the originality out of just about everything.

So naturally, I bring it up whenever I get the chance, as I love winding her up.

Most people have taken sides when it comes to artificial intelligence. Some love it and think it should be used for everything from giving you a recipe to writing reports at work.

Others, like my friend, are convinced it is stealing jobs and draining creativity by copying the work of writers, filmmakers and artists.

The reality is probably somewhere in the middle.

There is no question AI can make life easier. Nobody is going to miss spending hours on repetitive admin, data entry or similar jobs.

Have you noticed you are increasingly getting phone calls from AI trying to convince you to switch electricity providers or internet plans? “Jess AI” came up on my phone only the other day.

Used properly, it can make people more productive by summarising reports, transcribing interviews and handling tasks that would otherwise take hours, allowing them to spend more time on the parts that require creativity, judgment and human interaction.

But what happens when AI takes over the jobs where people learn their trade in the first place?

Last week, I gave a talk about journalism to a group of Year 10 students, when one of the English teachers asked me whether AI would eventually replace journalists.

It was a good question, as journalism is increasingly listed as one of the professions most exposed to AI.

A 2025 World Economic Forum Future of Jobs Report found 41 per cent of employers expect AI to reduce workforce numbers over the next five years, with a 2026 Global Talent Trends survey by global consulting firm Mercer finding entry-level and junior roles particularly exposed.

As well, Microsoft researchers analysed 200,000 real-world AI conversations and ranked journalists, writers, translators and customer service workers among the occupations with the greatest overlap between their work and what AI can do.

It found nurses, aged care workers, electricians, plumbers, mechanics, construction workers and roofers were among the occupations with the lowest overlap with current AI capabilities.

But research doesn’t always tell you the full story.

Can AI write a decent news story from a press release in four seconds? Absolutely. It can produce a headline, a standfirst, and an article with background faster than humans can.

But journalism, and writing books, films, poetry and the rest, is not just typing words into a document.

AI cannot investigate the story behind the press release. It cannot sit across from a politician and recognise the moment they are dodging a question. It cannot develop sources, build trust, read a room or know when something “doesn’t sound right”.

Prime Minister Anthony Albanese has limited journalists to one question at some press conferences. This allows him to give a polished spin answer without facing any harder follow-up questions that can often reveal the real story.

That requires judgment, instinct and experience, which is precisely why human journalists still matter.

As large corporations are finding out, AI can perform administrative and repetitive “grunt work” faster and more cheaply than people can.

But we also need to think carefully about what happens to entry-level graduate jobs if AI takes over.

Take paralegals. AI can already perform many of the tasks junior legal staff handle, such as reviewing documents, summarising cases and sorting research material.

But if machines do the junior work, how exactly are future lawyers trained? How do they learn through their mistakes at a low level when those mistakes don’t have a huge impact?

Many white-collar careers are built on a ladder, with graduates spending years doing routine work before progressing to more senior roles.

Junior accountants reconcile accounts and prepare tax returns before becoming finance leaders.

Graduate software developers start with bug fixes, testing and basic coding before designing complex systems.

Young marketers write social media posts, draft emails and conduct research before leading major campaigns.

See the common thread? Many of these entry-level tasks involve reading, writing, sorting and processing information, which is exactly where AI is becoming most capable.

Young workers traditionally learn through repetition, mistakes, observation and experience.

Over time, that becomes judgment, the one thing AI does not truly possess.

So how do you create experienced professionals if nobody is allowed to become inexperienced ones first?

Schools and universities are grappling with the rise of AI among students, with many using it for both research and actual writing. Most teachers use AI detectors for every piece of work they mark.

Businesses, forced to cut costs because of the state of the economy (thanks Albo), are increasingly relying on AI because they think it can do everything.

But do people really think information processing can replace human judgment and originality?

Eventually, governments may need to subsidise graduate jobs, and businesses will have to find new ways to train young workers.

When I looked at the bright and enthusiastic Year 10 students last week, I realised that conversation is barely happening, a huge part of the AI debate we should be talking about more.