AI Can Do the Work. It Cannot Do Your Formation

So a few years ago I took my kayak up to the Al Ain Adventure whitewater centre. It had been about a decade since I was last in my boat properly, and particularly on whitewater, and I fully expected to be the guy who would really struggle - who once was the expert but now was no longer. What surprised me was that I could do all of it. I don't even remember thinking about it. The use of the edge. The brace. How the hips move and the paddle follows through the water. None of it was lost. What was lost was my conditioning. The next morning my body felt it, painfully.

You see, the thing is, I've actually been trained in kayaking. I spent hours both on Loch Linnhe in the freezing cold water, with an instructor on the bank shouting at me at times, and on the rivers of the Scottish Highlands, learning and practising. And then later working as an outdoor pursuits instructor, week after week in water, being responsible for the safety and growth of others, with the skills and abilities being grooved into me as muscle memory.

This is something I think we can all relate to, in a different context. You probably drive a car, and you learned to do so by taking driving lessons. But the driving lessons, and even passing the driving test, didn't make you the driver you are today. They only really gave you just enough ability to be safe on the roads. It was the years of driving after that which gave you the unbreakable abilities you have now.

Meanwhile, something is happening to the entry-level job. In the most AI-exposed occupations, employment for 22- to 25-year-olds is falling, whereas it is holding steady or growing for older workers in the same jobs. Specifically, at large technology firms, the percentage of new hires who were new graduates fell by more than half between 2019 and 2024. That comes from the Stanford Digital Economy Lab, working from payroll data, and I only came across it while researching for this article.

So what's happening here? This is not the short-term market reaction of laying people off because of AI that we've probably all been reading about. This is about hiring that simply did not happen, because the machines were already doing the work.

I got both the training and the on-the-job formation, and that's partly why a decade later, when I get back in a boat, I've still got all the skills I had, having given them no attention at all. The new graduates who do get those entry-level jobs are only getting the former.

They'll be getting work done and producing quality output. The problems that previously would have slowed them down will quietly fall away, and another will take their place. They won't be going through the effortful process of forming what's needed later at higher levels of seniority. And it's not like someone senior made the conscious choice to take it away from them. To the entry-level employee starting their new job, and to the senior executive making the decision, it will seem like the obvious thing to do.

So how can we understand this in our current times? Mark Twain is said to have observed that history doesn't repeat itself, but it does often rhyme. When we think back to the Industrial Revolution we can see two things happening at the same time.

At first, something is automated, resulting in the displacement of formation that used to come with it.

What got automated was physical work and, as a result, formation of the body had to be found outside of work. The answer took the best part of a century: school physical education, organised sport, and eventually the slightly absurd business of exercising on purpose.

Following this, demand accelerates somewhere else.

Machines needed running and factories needed managing, so the premium moved to people with more education. Industries adopting the new technology employed more educated workers and paid them more. Secondary schools were built en masse, deliberately and at public expense, to supply them.

The underlying mechanism isn't new. There's a paper from more than forty years ago that describes the impact of automation on industrial control rooms. In short, the more something got automated, the more it pushed the need for employees to solve only the difficult edge cases. However, it was also noted that it was the work that got automated that had previously shaped employees into being able to solve the harder cases. The study was on power stations, but we can see the parallel with today's knowledge work being automated by AI.

What formation does AI displace, and what demand is being accelerated?

What's being displaced now is the formation of the mind. This happens when AI output is accepted indiscriminately, not challenged. It will show up in competencies like generating ideas from scratch, having an opinion and being able to stand behind it, making decisions in the midst of uncertainty, and being willing and able to openly disagree and be pushed back against. Previous generations learned this through years of attention and repetition. But what now?

As for the acceleration, we're seeing this already. AI does the mental work but needs direction. Someone has to be answerable for what it produces. Competencies here include specifying intent precisely, judging output you did not create, knowing when to override, and carrying the consequences either way. Work that previously only managers had to hold.

What does this mean for the way forward?

I don't purport to have the solution, but I can offer some thinking. We already know its shape. The body's answer took a century and ended up as exercise on purpose - things we do deliberately, outside of the work, because the work no longer does them to us.

All the competencies given above share one thing in common - they require being exposed. Vulnerability is, at essence, a distinctly human thing. AI isn't vulnerable. It's also been said that we might impress people with our strengths, but we connect with people through our weaknesses.

North Point Academy names Discipline of Growth as one of its values, and describes maturity as developing through sustained practice, reflection and challenge. This looks like deliberate vulnerability that forms us as humans, so that we can inhabit the roles and responsibilities that the future will ask of us.

Three practices to support us

From years as a coach, and training others as such, there are three practices that I can offer, and they all stem from the classic Johari Window that you probably already know. Each one moves something into better view.

1. Surface

This means bringing something to light from within that was previously outside of our awareness, or was there and unarticulated. This could come from journaling, painting, designing or working a problem from nothing, but also from reflection, prayer and meditation. AI is great at pattern matching, but it cannot have your insights for you.

2. Reveal

Telling others your previously unshared thoughts and experiences. Showing the parts of yourself that you've previously kept hidden. In doing so you do two things. Firstly, you show others that you're a person who can be trusted, and someone they can also open up to in return.

Secondly, you determine what you can be held to. Until it's said, others have only your behaviour to interpret, and they cannot know what you meant by it.

3. Receive

There's much that others know about us that we don't know about ourselves. Being willing to receive feedback about how you actually land is difficult - even vulnerable. It's also often the only way to see what we hadn't been seeing.

Why this matters now

When the output can be generated quickly, easily and at a higher level of quality than a given human can produce on their own, then trust is no longer built by the output. Instead, trust is built by having been wrong and been seen to handle it well, in relationship with others. Vulnerability is not a nice-to-have here. It's the only way left that people have to build trust.

And you?

1. Where am I choosing to lean on AI for the immediate result, knowing it's at the expense of longer-term formation - for myself, or for my team?
2. Where do I need to give attention to the skills needed to flourish in this AI age?
3. Where am I choosing to remain invulnerable, and thus missing the opportunity to connect with others?

In the NPA Certificate of Professional Coaching, we draw on the imagery of Plato's Cave to help us understand that you cannot support the formation of another person past the point you have been formed yourself.

Where is your formation calling for you to steward the formation of others?