This is the stance I get asked about with the most fear — and the one I feel the most fear about myself, because I am raising children into this too. So let me start with the reframe that dissolves most of it, the one I have to repeat to myself. The old job of a parent was to hand your child a map: study hard, get into university, get a job, be safe. That map is gone — not because the world got worse, but because it got faster than any map. The new job is to hand them a compass, and teach them to walk. My anxiety, and yours, is love looking for a plan. Here is the plan I am trying to follow.

First, the honest data — with the trap in it

You have probably heard that degrees no longer matter, and it is half true. Skills-based hiring is now the norm: some 85% of employers use it, only about 18% of job postings still require a degree, and McKinsey finds that hiring for skills is roughly five times more predictive of performance than hiring for education. But I refuse to over-correct, because there is a trap in the numbers: fewer than one in seven hundred new hires last year were people without a bachelor's degree. The rhetoric changed far faster than the practice. The degree still works as a filter and a signal, even where it is officially “not required.” So the honest counsel I give — and the one I hold for my own kids — is not “skip university.” It is: keep the degree a choice, not a foreclosed door and not an automatic default — and invest in the thing underneath it that makes any path work.

What I am actually trying to cultivate (durable, not vocational bets)

I do not try to pick my nine-year-old's career for a labour market fifteen years out that literally no one can forecast. I am trying to raise a human who can meet any future:

• Learning-to-learn. The single most future-proof skill. In a world where all content is free and AI tutors are infinite, the bottleneck is not access — it is self-direction, curiosity, and finishing what you start. The World Economic Forum expects 39% of core skills to change by 2030; the only constant is re-learning. I protect boredom and self-driven projects over a schedule optimised to the minute.
• Judgment and the right question. AI gives answers; the scarce skill is knowing what is worth asking and whether an answer is any good. I try to teach the reflex “how do you know that's true?” — the exact thing, the research notes, that AI tutors cannot teach.
• The deeply human. Empathy, real friendship, resilience, integrity, the courage to start things and the discipline to finish them. As the technical becomes cheap, these rise in value — the “human premium.”
• Depth plus adaptability. I want them to go deep enough on something they love to taste excellence — that discipline transfers and is where meaning lives — while staying broad enough to move.
• Making things. In a portfolio world, what they can do and who they are increasingly beat the diploma. I help them build, ship, and finish — a project, a piece of art, a small business, a garden.

The one insight the whole first series was pointing at

Here is the parenting move that matters most to me, and almost no one is making it: raise them to use AI as a superpower and to keep their core faculties in training without it. Educators are already watching children “offload their learning responsibilities” to AI, worried that too much assistance removes the productive struggle that long-term learning depends on. They are right, and I see it. So the rule in my house is simple: let them struggle with the hard thing before they reach for the machine. The struggle builds the muscle; the tool, used first, prevents it from ever forming. The child who can still think, write, and reason unassisted — and chooses when to use AI — will have a compounding advantage precisely because most of their peers will not. I will not let the cape wear my kids while their minds are still being made.

The safety part — and what it's really telling you

Two guardrails I hold to. First, AI literacy as safety: a child who understands that AI learns from patterns in data — and therefore can be biased, wrong, or manipulative — questions the machine instead of trusting it blindly, spots deepfakes, and guards their own digital footprint. I teach how it works, not just how to use it. Second, and gentler: about a third of teenagers now use AI companions, and the honest reading is that when a child turns to a chatbot for company, it is usually a signal about their social world, not the cause of a problem. The beloved-machine risk from the first series lands hardest on a mind still learning how to be with people. I watch for it — not to ban the tool, but to make sure the messy, difficult, irreplaceable human bonds that actually build a self are still there and still winning.

And how you should feel

Here is the reassurance I keep giving other parents, and myself: at these ages, the highest-return investment is not the right vocational bet. It is the human core — curiosity, confidence, kindness, resilience, the ability to self-teach, a sense of what they love and that they are loved. Those never go obsolete. And the single strongest predictor in the research is not a curriculum or an app; it is you. Children whose parents engage with AI learning alongside them show stronger critical thinking about it and more responsible use. You do not need to know more about AI than your kids do. You need to model the one thing you already have: a person who keeps learning, keeps their judgment, and is not owned by their tools. Children learn the compass by watching a parent use one — which means the work starts with me, not with them.

Three honest cautions I keep near to stay level. Do not foreclose university prematurely — remember the one-in-seven-hundred. Do not forget that self-education needs a scaffold: a stable, curious, supportive home, which is itself an advantage not every child has, and which your engagement supplies more than any program could. And do not turn childhood into a skills-acquisition project — protect the play, the boredom, the friendships. The “human premium” grows in exactly that unoptimised soil.

I cannot hand my children a map for a world I cannot predict. But I can try to raise the kind of person who thrives in unpredictability: curious, self-teaching, kind, resilient, and quietly disciplined about the mirror. Give them the compass. Model the walking. Then trust them to steer. That is not a lesser gift than the map. In this world, it is the only one that still works — and it is the one I am trying, imperfectly, to give.