AI is changing how people work. Is it also changing how they learn?
Created with
One of the most exciting promises of AI is that it can free people from some of the more repetitive and time-consuming parts of their jobs.
Ask AI to summarise the research. Produce the first draft. Analyse the data. Turn the meeting into actions. Pull together the initial recommendations.
That can make work faster and, used well, create space for people to focus on more valuable activity. But there's an important employee experience question hiding inside that opportunity: What if the task AI just saved someone from doing was also helping them learn?
A lot of professional development happens without us necessarily describing it as learning.
Junior employees research things they don't yet fully understand. They attempt first drafts. They sit in meetings and listen to experienced colleagues navigate difficult conversations. They analyse information, make recommendations, get things wrong and try again.
Gradually, they develop judgement.
This is particularly important because expertise isn't simply knowing more information. It includes recognising patterns, understanding context, knowing when the usual answer doesn't apply and developing the confidence to challenge something that doesn't look right.
A 2026 paper describes the potential loss of these developmental experiences as the "apprenticeship void".
As AI absorbs more foundational knowledge work, the researchers argue, people may have fewer opportunities to build expertise through actually doing that work.
And that's something EX practitioners should be paying attention to.
Imagine someone early in their career using AI to produce an excellent first draft. They might get to a better output much faster than they could have produced independently. That's valuable.
But what have they learned along the way? It depends entirely on how the interaction has been designed.
If AI produces the answer and the employee accepts it, very little learning may have happened. If the employee has to interrogate the answer, understand its reasoning, identify weaknesses, improve it and explain the choices they've made, the same technology could become part of the learning process.
That's an important distinction.
The question isn't whether junior employees should use AI. They absolutely should be learning how to work effectively with it. The question is whether we're designing AI-enabled work purely around getting to the output faster, or whether we're also thinking about the capabilities people need to develop along the way.
There's another dimension to this.
As AI becomes better at technical and routine cognitive work, we're hearing more about the growing importance of distinctly human capabilities: judgement, communication, empathy, creativity, collaboration and critical thinking.
Recent reporting from professional services suggests some firms are putting renewed emphasis on face-to-face experiences for junior employees partly because of concerns about how those skills develop as AI takes on more routine work.
It's an interesting reminder that we can't simply declare human skills more valuable and expect people to acquire them.
If anything, we may need to create more deliberate opportunities for people to observe, practise, discuss, question and receive feedback.
For EX practitioners, there is an opportunity to bring a different perspective to conversations about AI adoption.
Rather than only mapping which tasks AI could automate, we could also explore the developmental value hidden within those tasks.
Where do people currently learn by doing? Which experiences help someone move from novice to expert? Where do employees develop judgement rather than simply knowledge? What opportunities exist to observe experienced colleagues? And if AI removes a task, what needs to replace the learning that came with it?
This doesn't mean protecting repetitive work for the sake of it. It might mean redesigning junior roles so people critique AI-generated work rather than simply receiving it. It could mean deliberately building reflection and feedback into AI-enabled workflows. It might mean creating more opportunities for shadowing, mentoring, problem solving or explaining the reasoning behind decisions.
The aim isn't to preserve the old way of working. It's to make sure we're not accidentally designing the learning out of the new one.
Because AI could give us an extraordinary opportunity to rethink early-career work. We can remove some of the drudgery while creating richer, more intentional ways for people to develop.
But that won't happen automatically.
As EX practitioners, perhaps one of our most important questions as AI changes work is becoming: what experiences will people need today to become the experts we'll need tomorrow?