Some stories in human history do more than reflect their own time. They continue to speak to generations that follow because they capture something enduring about human nature.
One such story is The Sorcerer’s Apprentice.
The apprentice brings a broom to life and orders it to carry water, hoping to make his work easier. The broom neither tires nor complains. Delighted, the apprentice believes he has freed himself from repetitive labour.
But that is where the problem begins. He has learned how to make the broom start, but not how to make it stop. The water keeps rising, the situation slips beyond his control, and every attempt to stop the broom only makes things worse. In the end, the power he has created proves greater than his ability to govern it.
This centuries-old story has an unexpectedly close connection to today’s debate about artificial intelligence.
Every transformative technology arrives with two competing narratives.
The first is one of hope. AI could accelerate medical breakthroughs, compressing decades of research into years. It could help teachers personalise education, help farmers increase yields and help doctors identify illnesses earlier.
The second is one of fear. AI could displace millions of jobs, allow misinformation to spread at unprecedented scale, centralise power in the hands of a few and eventually produce systems whose behaviour we do not fully understand.
There is truth in both narratives.
The printing press democratised knowledge. Electricity expanded productive capacity. The internet connected billions of people. Every transformative technology has reshaped society while creating new problems. AI will probably be no different.
A few weeks ago, I had the opportunity to use Claude AI at work. Tasks that had sat on a “one day” list for years were completed within two weeks. Work I had expected to take months moved much faster. Research accelerated, drafts improved and analysis became more structured.
My first reaction was amazement. I had not imagined so much could be done with so little time and effort. But the excitement did not last.
I had assumed that becoming more productive would mean I could finally sit back. Instead, I found myself asking different questions.
Productivity had risen, but so had a subtle unease. If AI can compress months of work into weeks, what is my real contribution? Is expertise about possessing information, or about knowing what to do with it? Is value created by producing an answer, or by making the right decision in the right context?
These are not merely personal questions. Millions of knowledge workers may be wrestling with the same uncertainty.
For much of human history, knowledge was scarce. Those who possessed more of it held greater influence. Expertise was limited, research was slow, and gathering information was itself a competitive advantage.
That world is changing. Today, anyone with access to generative AI can retrieve, organise and synthesise the equivalent of thousands of pages within minutes. The question is no longer simply who has information. The question is where value lies when information becomes abundant.
There is a related challenge. When generative AI is this accessible, almost anyone can sound like an expert. A few well-phrased questions can produce answers that appear authoritative, confident and complete.
But AI answers the question it is asked. It does not necessarily know the situation behind the question. It does not know why a decision matters, who it will affect or which constraints were never included in the prompt. It can produce a polished answer without knowing whether that answer actually fits the context.
The danger, therefore, is not simply that people will stop knowing things. It is that fluency will be mistaken for understanding.
A calculator can perform mathematics, but it cannot make life decisions. GPS can show a route, but it cannot choose the destination. AI can present options, but the responsibility for choosing between them remains human.
That may make judgement one of the most valuable skills of the AI age.
The ability to ask the right questions in complex situations, make decisions despite incomplete information, understand human context and recognise consequences may matter more as machines become better at producing information and possible answers.
This shift is not limited to individuals. It is already playing out at the level of organisations.
Every organisation is asking the same basic question: will AI make us more effective? The answer, in many areas, is clearly yes.
Reports, code, analysis, legal drafting and customer service can increasingly be produced faster and at lower cost. If one employee can use AI to accomplish work that previously required several people, organisations will have strong incentives to rethink how work is organised.
The change may not happen overnight. It could begin with AI acting as an assistant, making individual workers more productive. Over time, however, organisations may ask whether they still need the same number of employees, whether vacancies need to be filled, whether teams can be smaller and whether some processes can operate with minimal human involvement.
The outcome is not predetermined. Technology has repeatedly eliminated some forms of work while creating others. Telephone operators largely disappeared, while software engineers emerged. Typewriter specialists faded, while the digital economy created entirely new professions.
Something similar may happen with AI, although its scale and speed could make the transition different from previous technological shifts.
The implications extend beyond individual jobs and organisations. The same capabilities are now shaping how nations compete.
AI itself is neither inherently good nor bad. Fire can cook food or destroy a home. The same technological capabilities that help protect a hospital can also be used to identify vulnerabilities. Systems that strengthen national infrastructure can also expose ways to compromise it.
AI is therefore no longer merely a commercial asset. It is increasingly a strategic one.
A technology that can accelerate scientific research, strengthen cybersecurity, increase economic productivity and support national decision-making can influence economic, scientific and geopolitical power for decades to come.
The United States, China, Europe and countries in the Middle East are all investing heavily in AI. That creates another dilemma. “Let’s stop and think” may sound like the obvious response to a technology this powerful, but it is difficult to put into practice when companies and countries fear that moving more slowly could leave them behind.
The result is a familiar tension. The greater the potential of AI becomes, the stronger the incentive to develop it. And the stronger the incentive to develop it, the more important it becomes to consider how that power will be governed.
Does more intelligence automatically mean a better future? Perhaps not.
AI can analyse information, solve complex problems and operate at unprecedented speed. But intelligence does not automatically provide compassion, morality or purpose. A highly capable system can pursue the wrong objective just as efficiently as the right one.
Intelligence and wisdom are not the same thing.
That may be one of the most important lessons the AI age teaches us.
In The Sorcerer’s Apprentice, the problem was never the broom itself. Nor was it simply the power the apprentice had discovered.
The problem began when his ability to create power outgrew his ability to control it. That is where the real question about AI begins.
Not simply whether we can build systems that are more capable, faster and more intelligent, but whether our ability to govern them, understand their consequences and exercise judgement can keep pace.
The challenge of the AI age may therefore be less about whether machines can produce more answers.
It may be about whether humans can continue to ask the right questions. In the age of AI, information is no longer scarce. Clear thinking is.
