Yes, I know how that sounds. Cold. Transactional. The kind of thing an economist would say. I am an economist, and a parent, and I am going to argue this is the warmest position available. Hear me out.
This week I co-hosted a roundtable at SXSW London alongside Anita Cleare on “AI readiness starts at age zero”. Around the table were parents, educators, HR directors, and policy folk, which made for exactly the kind of cross-disciplinary friction this topic needs. This is the case I made to them.
Anita and I hanging out at the speaker’s lounge beforehand.
The useless binary
Many people are increasingly seeing that the binary division into AI optimists and pessimists is completely useless. The techno-optimists tell us that AI will bring a world of abundance whereas the techno-pessimists warn us about the destruction of humanity. Makes for excellent clickbait but it doesn’t leave you any wiser. In the debate around AI the optimist/pessimist binary is the equivalent of junk food.
How do we balance this out?
By taking a longer-term, systems-wide perspective.
The real issue we should be focusing on is what actions you take today and tomorrow to close the AI readiness gap.
The AI readiness gap that’s becoming a chasm
The AI readiness gap is the disconnect between what companies - and governments - are spending on their AI technology infrastructure on the one hand, and how prepared and confident people feel to use those tools. Only 17% of the UK public say they can explain AI in detail and 28% feel confident using AI at work.
Lack of confidence with AI is by no means just an employee issue. My own data from the research I’ve conducted for my forthcoming book “AI-Q” shows this. When I asked 200 leaders how confident they felt about leading their teams through AI-driven change just 31% of them report feeling very confident.
It is no surprise that words like techno-stress, AI-slop and brain-fry are now commonplace in our daily vocabulary. At another conference I spoke at this week, KIMRA, (yes June seems full of events!) I frequently heard people use the word “overwhelmed”. The audience was professionals from the research, insights, knowledge management and thought leadership functions across banking, legal firms and consulting.
There seemed to be two camps.
Those that felt giving us all a powerful tool doesn’t free up time, it instead expands the range of possibilities available to us, and we end up wanting to do more. All of a sudden you feel that you can do more, so you should, making it harder to switch off. With one attendee we talked about what this might mean for future health insurance claims down the line, but I am diverting away from the topic at hand.
The other camp was those that felt overwhelmed by the pace of change, and the renewed expectations by their stakeholders to deliver more and faster than ever before, and in some cases that led to a freeze response. For a function like research and insights where quality is paramount, that acceleration is hard to adapt to.
Why does all this matter? Because much of the conversation around AI still focuses on which tool is better, chatGPT or Claude, what is the mother of all prompts that will avoid hallucinations (er, none), and the different cost of tokens. When you talk to real people you realise the complexity of what is at hand.
Canaries in the coalmine
Today’s graduates are the canaries in the coalmine. The eponymous research paper by Erik Brynjolfsson, Bharat Chandar, Ruyu Chen find that:
“since the widespread adoption of generative AI, early-career workers (ages 22-25) in the most AI-exposed occupations have experienced a 16 percent relative decline in employment even after controlling for firm-level shocks. In contrast, employment for workers in less exposed fields and more experienced workers in the same occupations has remained stable or continued to grow.”
I see this on the ground and first hand through my work with BossUp.AI, where I get to speak to many young graduates here in the UK, and they tell me they feel let down by the education system’s failure to prepare them for a world of work, or at least the recruitment aspect and the skills they will be asked to prove in order to get a job. Some tell me they apply for 500 jobs without a response and how they find that “soul destroying”. Some wonder whether they should give up, or ask me what they are doing wrong. Is it any surprise that in the UK alone we have more than 1 million young people that are not in education, employment or training (NEETs)?
They see the headlines. Every week brings another story about AI replacing entry-level jobs, another CEO boasting about hiring freezes. Imagine being 22, applying for your five hundredth job, and reading that. The message young people receive is that the economy is being redesigned without them in it. No wonder that recent speakers at US graduation ceremonies have been booed at the mere mention of the word “AI”.
The AI skills paradox
I’ve been asking executive after executive that I interview or speak to “which skill, or ability, will be the most important in the future?”
They consistently say that the skills they will need, and the skills they will value the most are:
critical thinking
empathy
strategic thinking
creativity
learning agility
These abilities will be the biggest differentiators in a workplace where human and artificial intelligence co-exist, and what they all have in common is that they are deeply human abilities.
The irony lives in the fact that those same executives are investing in AI tools1 that are undermining the pipeline of these skills for the future. In many cases, whether it’s because that is the way the AI tools have been designed or because of the way we choose to use them, they end up removing the mechanisms through which these abilities are developed. How can you develop your critical thinking skills when AI tools are designed to bypass that altogether?
Here it’s essential to distinguish between someone that has developed a particular ability over the years, and that ability atrophies, and someone who never develops that ability in the first place, such as our children and young people. This is why the parenting/carer angle matters so much.
The parenting angle
Here are the provocations my co-host Anita Cleare shared with the group:
Lack of time + cultural pressure to optimise children leads towards solutions that promise big learning gains for minimum adult input. Many parents see tech skills as a route to helping kids get ahead.
Young children build their brains and human skills through real world play and inter-personal interaction. For example, the brain develops differently when you learn by touching something physical or by putting pen to paper, compared with reading something on a screen.
The children who will thrive in an AI age won’t be the ones who use AI earliest. This chimes with what I hear from executives: the skills that will be needed most in the future are the human skills, and those are built away from the screen, in the playground, around the kitchen table, in the messy business of dealing with other people.
She also shared a stat that in some parts of the UK it is reported that 38% of 10 to 18-year-olds find it easier to talk to AI than to their parents.
The case for treating childhood as economic infrastructure
All this begs the question, what do we do about it?
My argument is that we need to start treating childhood as economic infrastructure. Preparing the future generation for the world of work should not start when they are 16 and approaching their final studies, or after they’ve graduated. It needs to start much earlier, hence “age zero”. It means that we need to start realising young people are the most essential diffusion mechanism.
One criticism I have had is that this is a rather cold view of seeing our children, and that it commoditises childhood. I would argue the opposite. Economic incentives right now are geared towards:
eliminating friction
speeding up teaching
rewarding instant outputs over slow skill-building
But once we start seeing childhood as economic infrastructure, the logic flips. We do not commoditise our roads, energy grids or broadband networks by calling them infrastructure. We protect them. We invest in them for the long term, we maintain them, we hold ourselves accountable when they degrade, and we treat underinvestment as a national risk rather than a private problem.
What if we apply that lens to childhood?
Three things change.
First, responsibility shifts. The development of critical thinking, empathy and creativity stops being something we outsource to individual parents juggling impossible time pressures, and becomes a shared obligation across employers, educators, technology companies and government.
Second, time horizons shift. Infrastructure is planned in decades, not quarters. The executives investing billions in AI tools today would be forced to ask what those same tools are doing to the talent pipeline they will depend on in 2040.
Third, design shifts. If childhood is infrastructure, then friction is not always a bug or problem to be eliminated. Some friction, the kind involved in wrestling with a hard problem, learning to read another person’s face, or putting pen to paper, is precisely how the “asset” gets built. Tools designed for children, and for the adults raising them, would need to protect that productive friction rather than engineer it away.
Far from being a cold view, this is the warmest argument available. It says that what happens in a sandpit, around a dinner table, or in an unhurried conversation between a parent and child is not a soft, private nicety. It is the foundation of every deeply human ability those executives tell me they will pay a premium for. The coldness lies in the current arrangement, where we let market incentives quietly strip-mine those foundations and then express surprise when graduates arrive at the labour market underprepared and overwhelmed.
What you can change on Monday morning
For parents and carers: you do not need to teach your five-year-old to prompt. Prioritise the real-world play, conversation and boredom through which the brain builds the abilities AI cannot replicate. As Anita said in her commentary: “The children who thrive in the AI age will not be the earliest adopters, they will be the ones with the strongest human foundations.”
For employers: stop treating entry-level roles as a cost to be automated away and start treating them as the apprenticeship system for your future leadership. Examine how those entry level workers could be the very diffusion mechanism that your AI investments need to scale across the organization. If the bottom rung of the ladder disappears, so does everyone who would have climbed it. Audit what your AI deployments are doing to skill formation, not just to productivity.
For policymakers: the AI readiness gap will not be closed by infrastructure spending alone. If we can publish league tables for school buildings and broadband rollout, we can measure and invest in the development of human capabilities from the earliest years, and we can hold the system accountable for the one million young people currently locked out of it.
The optimists and pessimists will keep shouting at each other. Meanwhile the readiness gap widens, the canaries keep singing, and the abilities we say we value most are quietly going underfunded at the exact stage of life when they are built. Treating childhood as economic infrastructure is not about turning children into assets. It is about finally giving the foundations of human ability the seriousness, the investment and the protection we already give to everything else we cannot afford to lose.
I refer to the suite of generative and agentic AI tools, rather than AI more broadly, which also includes rules-based AI, machine learning, deep learning and physical AI.
It is economic infrastructure, this has been the case since ever. The key point is to be treated as one. This would extend to the value society puts to carers. At the moment they perceived passive elements of the economy and usually there is no financial reward on the contrary usually consequence to one’s career and financial income. Which usually is woman. So your perspective I feel will address more than one problems