Schools and universities have an AI problem that companies should pay attention to.
Students can now hand in perfectly respectable essays and homework without necessarily having done much of the thinking themselves. If there is an easier way to get something done, most of us will use it.
But what happens to the learning?
Schools and universities will have to find new ways of testing whether somebody has actually understood something. Oral exams are already making a comeback.
Companies are facing a version of the same problem; except we tend to talk about it as productivity.
We are working out what AI and agents can take off people's desks. A lot of what they can take away is the work traditionally done by younger employees. But that work was also how they gained experience.
Judgement comes with experience. In the past, we developed it by sitting in client meetings and not yet understanding all the dynamics. Preparing something that a more experienced colleague took apart. Watching a negotiation take an unexpected turn. Over time, you learnt what to pay attention to and what to question.
The answer should not be to make young people do work that AI can do perfectly well just because we had to do it ourselves. We need to think about how they will learn when AI is part of how they work from day one.
They can use AI for the research, but they should be able to explain why they trust the result. They need to challenge an AI recommendation rather than just present it. And perhaps removing some of the more mundane work will give them the opportunity to get closer to customers and important decisions earlier in their careers.
There should probably also be times when AI stays switched off. Pilots have autopilots, but we still expect them to know how to fly the plane and respond in critical situations that the autopilot cannot handle.
The experienced people checking and challenging AI today won't be doing it forever. The next generation still needs to develop the judgement to take their place.
This needs to be part of how companies think about AI value creation. Which work can AI do, but also how will people learn, gain experience, and eventually take responsibility?
What does an apprenticeship look like when AI is doing much of the work apprentices used to do?