Which skills will help people most in an AI-driven future?

Norway's recent announcement that it was close to banning artificial intelligence (AI) in its elementary schools has sparked debate about the role of this technology in education. While most Norwegians have welcomed the move, others argue that delaying exposure could leave the country's students less prepared for an AI-driven future.

Other countries are taking a different approach. Poland, for example, is introducing AI labs into primary and secondary schools, while the UAE is bringing AI into the curriculum from kindergarten.

These different approaches raise questions: does preparing children for an AI-driven future have to mean getting them to use the technology as early as possible? What if some of the capabilities we are trying to automate are the ones that will become the most valuable in the AI era?

This focus on long-term capability over short-term gains is not unusual in Norway. It reflects a broader Nordic philosophy, perhaps most clearly seen in Norway’s approach to sport.

Norway’s youth sporting system is built around development rather than early competition, with a focus on idrettsglede, or the joy of sport. Physical activity, inclusion and fun in sports take priority over identifying talent and accelerating performance as early as possible.

This approach brought home 41 medals, 18 of them gold, at the most recent Winter Olympics. Months later, Norway’s men's football team reached its first World Cup quarter-final, defeating Brazil along the way.

Of course, there is no straight line between children enjoying sport and Olympic medals. But the underlying philosophy is don’t rush development in pursuit of an immediate result.

Norway is now applying a similar principle to AI use by resisting the pressure to accelerate simply because the technology allows it.

When it comes to business, that pressure is understandable. AI can produce work faster, analyse more data, generate ideas and take care of many of the mundane tasks that fill people’s working days. As organizations spend significant amounts on the technology that facilitates this, they understandably want to see a return.

But there’s a problem with assuming that more AI will automatically create material advantage.

Access to these capabilities is becoming more widespread, and the value proposition is no longer clear-cut. An organization’s competitors will have access to similar models, similar computing power and much of the same underlying intelligence. Having AI will not, in itself, make an organization different.

What could make it different is the quality of the people using AI.

Judgement, critical thinking, creativity, curiosity and the ability to make connections between seemingly unrelated things will all matter in an AI-driven future. And none of these capabilities are suddenly acquired once a person reaches a certain level of authority. They are developed by doing the work.

People learn judgement by making some bad decisions. Critical thinking develops through wrestling with problems. At the heart of creativity is experimentation, failure, inspiration from the world around us and occasionally sticking with an idea when everyone else thinks it’s a bad one.

A lot of this is inefficient – and that’s partly the point. While AI can remove a lot of that friction, if it removes the experiences that help people learn, today’s workforce may become more productive at the expense of the full development of tomorrow’s.

Coming back to Norway’s AI in schools policy, the interesting question isn’t really whether a 10 year-old should be using ChatGPT, it’s what we want that 10 year-old to be good at by the time they enter the workforce.

That requires something that many businesses are not always very good at: patience.

Patience is, admittedly, a difficult sell to business leaders right now. They’re dealing with geopolitical uncertainty, economic pressure and technological upheaval. Nobody wants to be the company that waited while everyone else moved ahead.

But Norway isn’t rejecting technology or AI – and nor should businesses. At the crux of Norway’s AI in schools policy is the conundrum of what happens to people as more of the work they once learned from is handed over to machines.

Most of us developed expertise early in our careers by researching things we didn’t understand, writing bad first drafts, making mistakes and watching more experienced colleagues pull them apart before we tried again. A lot of that work can now be done by AI in seconds, rather than helping people to learn by doing.

If organizations automate too much of that experience, they risk creating a strange kind of productivity paradox, with more work being produced, faster than ever, while fewer people get the opportunity to develop the judgement or critical thinking needed to know whether the output is actually any good.

So, instead of asking how much technology is being deployed, leaders should be asking where this leaves the capabilities of their employees?

Are people getting more curious or more dependent on a machine? Are they gaining judgement or losing the friction that builds it? Organizations need to decide where AI should be used, where humans should remain actively involved and where a degree of productive struggle is worth protecting.

The most advanced organizations may not ultimately be those that automate the most, but those that are much more deliberate about what they automate in the first place.

Norway’s sporting model took decades to develop, with tweaks and changes along the way. It will also take years to see whether its approach to AI has been a success.

But this shouldn’t be seen as a binary choice between AI and humans. As access to powerful models becomes more widespread, judgement and originality will become more valuable. So will people who are willing to challenge an answer rather than simply accept what AI tells them.

The lesson from Norway is that developing these capabilities takes patience. The AI race won’t be won by those applying it the fastest, but by those who develop people who know how best to use AI once everyone deploys it.