Artificial intelligence is already changing the shape of work. Used well, it can remove friction, accelerate delivery, improve analysis and give capable people more time to focus on judgment, creativity and problem solving.
I am not against that. In many areas, I am excited by it.
But there is a risk in the way organisations adopt AI if they only look at the immediate efficiency gain. Some of the work most likely to be automated is the same work that once gave young people a way into the workforce. The junior analyst compiling notes. The graduate preparing the first draft. The assistant checking documents, coordinating basic tasks or sitting close enough to a process to learn how decisions are made.
Those tasks were not always glamorous. Many were repetitive, administrative or menial. But they were also apprenticeships in disguise.
They gave people exposure to context. They taught them how organisations really operate. They allowed them to see how senior people think, how decisions are framed, how mistakes are corrected and how small details connect to the bigger picture. A young person did not begin with full judgment. They built it by being close to the work.
If AI removes too much of that early exposure, we may solve one productivity problem while creating another. We may become more efficient in the short term but weaken the pipeline of people who understand how to operate, lead, govern and make decisions in the long term.
That matters because expertise is not only the final answer. It is the experience behind the answer. It is pattern recognition, context, restraint, ethical judgment and the ability to know when something that looks correct is actually wrong. Those capabilities are hard to develop if the first rung of the ladder disappears.
This is not just an employment issue. It is a social issue.
Young people entering the workforce need more than income. They need belonging, responsibility, feedback, mentors, frustration, small wins, difficult conversations and the chance to discover what they are good at. Work has always been one of the ways society teaches people how to contribute. If organisations use AI to strip out too many entry-level opportunities, we should not be surprised if the outcomes reach beyond the workplace.
Responsible AI adoption needs to recognise this.
Boards, CEOs and executives should not only ask, “How much cost can we remove?” They should also ask, “What learning pathway are we removing?” If a junior role is automated, where does the next generation learn the work? If the first draft is always produced by AI, who learns to write? If basic analysis is always automated, who learns how to question the data? If routine coordination disappears, who learns how people, systems and accountability actually connect?
The answer is not to reject AI. That would be unrealistic and, in many cases, irresponsible in its own way. The answer is to adopt it with intention.
Organisations should use AI to improve work while deliberately preserving opportunities for people to learn. That may mean redesigning graduate roles rather than deleting them. It may mean pairing young staff with AI tools but still requiring them to explain, challenge and improve the output. It may mean creating new apprenticeship models where people learn by reviewing, governing and shaping automated work instead of simply performing the old manual version of it.
It also means being honest about what AI is good at and what it is not. AI can produce a useful first pass. It can summarise, draft, search, compare and generate options. But it does not carry responsibility. It does not understand consequence in the human sense. It does not build the quiet confidence that comes from years of seeing how decisions affect people, customers, communities and organisations.
That responsibility still belongs to us.
The danger is not that AI becomes useful. The danger is that we treat usefulness as the only measure that matters. If we allow every efficiency decision to stand alone, we risk changing the way people enter adulthood, develop capability and find their place in the world.
AI can help us build better organisations. It can help people move faster, think differently and focus on higher-value work. But if we use it without care, we may also remove some of the messy, imperfect pathways that helped people become capable in the first place.
That is the balance we need to get right.
The future should not be a choice between productivity and human development. A responsible society should be able to pursue both.