The job your child will get hasn't been invented yet

A child in Grade 6 today will start their first serious job around 2033 or 2034. That's not far away in the way careers feel. It's one degree, one set of exams, one internship away. And the uncomfortable part is that the world they're walking into will not be a slightly upgraded version of ours. It will be structurally different.

I write and consult about AI and digital transformation for a living. I spend my days telling businesses to rebuild themselves around machines that think. And lately, when I look at my own kids, the same conversation comes back at me from the other direction: the businesses are ready to change faster than the school system is, and my children are sitting in the school system.

Here's what the evidence actually says, what the next decade of work will look like, and what a parent should do about it now.

The disruption isn't coming. It's scheduled.

Let's put some numbers on the table, because "the future is changing" is a slogan, and slogans don't help anyone plan.

The World Economic Forum's Future of Jobs Report, published in January 2025, projects that 170 million new jobs will be created and 92 million jobs displaced by 2030 — a net gain of 78 million, but a brutal churn underneath. About 39% of the core skills a job demands today will be different by 2030. That's not a footnote. That's nearly half of what you're hired to do.

Dell Technologies and the Institute for the Future projected back in 2017 that 85% of the jobs that today's learners will do in 2030 haven't been invented yet. Eight in ten. If you read that and think "well, that was a marketing slide," fine — but it's held up better than almost any forecast in tech, because every subsequent year has moved the goalposts earlier.

The rate of obsolescence is the quiet part. Gartner tracked that the half-life of a skill — the time before it's half as valuable as it was — fell from about 10.5 years in 2010 to about 4.5 years in 2022. For hard technical skills, it's shorter still. In your parents' working life, a skill you learned at 22 could carry you to retirement at 60. Your child will need to re-earn their edge every few years.

And the scale of people affected is not small. McKinsey's "Jobs lost, jobs gained" work estimated that automation could displace 400 to 800 million jobs globally by 2030, with up to 375 million people needing to switch occupational categories entirely. Goldman Sachs, in March 2023, estimated that generative AI alone could expose the equivalent of 300 million full-time jobs to automation.

Put those together and the picture is clear. This is not a wave lapping at the edges of the labour market. It is already breaking, and it will land right as today's Grade 6 students graduate.

What the new career will actually look like

The old model was linear and predictable: study hard, get the degree, land the job, climb the ladder, retire. It was already crumbling before AI. The next decade finishes it off.

The credential is losing its value as a filter. It used to be that a degree told an employer "this person can be taught." Machines changed that. When a tool can draft the document, analyse the spreadsheet, and write the first version of the code, the employer no longer pays you to recall information. They pay you to decide what the machine should do, to know when it's wrong, and to be trusted with responsibility.

So the future employee is less a repository of knowledge and more a manager of intelligence — human and artificial, at the same time. The jobs that grow will be the ones that pair judgement with a machine. The jobs that shrink will be the ones that were mostly inputs, data entry, routine reports, standard customer replies, first drafts of anything.

Careers themselves will become portfolios, not ladders. People will hold several roles at once, switch fields every few years, and be hired for demonstrated ability rather than a stamped piece of paper. That's already visible in the freelance and creator economy, and it will keep spreading up the corporate chain.

The soft skills that will decide who thrives

This is where I want parents to stop worrying and pay attention, because the good news is that the most valuable things your child can build are not machines. They are the things machines cannot do well.

Judgement and taste. The machine gives you ten options. Knowing which one is right, and why, is now the highest-value human act. This comes from reading widely, arguing, making decisions and living with the consequences.

Asking better questions. Answers are free and instant. Questions are the scarce resource. A child who learns to question assumptions, reframe problems, and chase "why" will be worth more than a child who memorised every formula.

Communication that changes minds. When every written first draft is AI-generated, the premium moves to persuasion, to storytelling, to explaining a hard idea so clearly that people act on it. This is a craft, and it is badly under-taught.

Adaptability, or comfort with being wrong. The half-life problem means your child's hard skills will expire, repeatedly. The people who survive treat learning as a habit, not an emergency.

Empathy and trust. Machines can be intelligent. They are not accountable, and they cannot hold a relationship. Leadership, negotiation, and care work all grow in value precisely because they stay human.

Here's the thing that surprises people. Our exam system — the board exams, the coaching classes, the marks that decide a stream in Grade 9 — rewards exactly the wrong things for this future. It rewards speed, memory, and reproducing a known answer under time pressure. The future pays for slow thinking, original thinking, and comfort with questions that have no known answer. There's a mismatch between how our children are judged and how they will be paid.

What to do about it now, not later

You don't need to throw out school. You need to stop treating it as the whole job. A few practical shifts, age by age.

Build a habit of making things, not just consuming. Whatever it is — a video, a small project, a story, a piece of code — the child who makes things learns iteration, feedback, and resilience. Let them fail at it early and cheaply. This is the closest thing to "upskilling" that actually scales.

Teach them to use the new tools as amplifiers, early and openly. The current instinct is to ban AI in schools. That instinct will hurt these children. The students who learn to use these tools well — to direct them, to check their work, to know their limits — will graduate ahead of the ones who were protected from them. This is the biggest single advantage a parent can give right now.

Push real projects over more past papers. An internship, a volunteer role, a family business task, a science fair project that actually fails once. Real-world problems teach judgement in a way no worksheet can. In Bangladesh, in India, this matters doubly, because the coaching culture leans so heavily toward exam output that real projects feel like a luxury. They're not. They're the point.

Read, write, and argue every single day. Not for the marks. For the judgement, the persuasion, the taste. This is the cheapest, highest-return investment in the whole list.

Model adaptability in your own life. Kids copy what they see. If you're visibly learning new things, changing your mind, and admitting mistakes, they'll treat those as normal and safe. That's the deepest soft skill there is.

The reassurance that matters

I want to be honest about the fear, because I feel it too. It is genuinely unsettling to tell a 12-year-old "the thing you're being graded on may not be the thing you're paid for." But fear is a bad planner, and the evidence cuts both ways.

Every major technology wave destroyed jobs and created more, and this one is creating enormous numbers of them — the WEF projects 170 million new roles. The human skills at the centre of this transition — judgement, communication, trust, adaptability — are the skills humans have always had to develop the hard way. Nobody can take those from your child, because they aren't storable in a machine.

So here's my advice, stripped down. Stop asking "which stream or degree will be safe." There is no safe degree. Start asking "is my child becoming someone who can learn, decide, persuade, and adapt?" Because that person will be fine, whatever the machines learn to do next. Start today. A 12-year-old has a decade to build these habits, and that's a genuinely enormous head start.