Synopsis
For years, the debate around artificial intelligence has been framed as a question of replacement. Will AI take our jobs? Will AI become more intelligent than us? Will machines eventually make human beings unnecessary?
I increasingly believe that this framing misses the more immediate problem. The greatest danger may not be that artificial intelligence becomes too powerful. It may be that human beings become too comfortable relying on it.
What concerns me today is not a distant science-fiction future in which machines suddenly seize control. What concerns me is what I can already see in ordinary business life: websites that look almost identical, presentations that use the same language, founders who describe entirely different products with the same vocabulary, and companies that appear to be outsourcing not only production, but judgement.
Artificial intelligence was supposed to increase our ability to create, analyse and execute. Yet if we are not careful, it may also encourage a generation of professionals to stop questioning, stop editing and eventually stop forming their own point of view.
The problem is therefore not simply whether AI will replace humans. The more uncomfortable question is whether humans are slowly learning to replace their own thinking.
I Began Noticing It in Websites
I have spent a large part of my professional life building websites, software systems, digital platforms and businesses. Because of that background, when I look at a website, I rarely judge it only by whether it looks modern, clean or attractive. I instinctively look at the thinking behind it. I notice how information has been prioritised, how the company explains itself, whether the customer journey makes sense, whether the wording reflects a genuine understanding of the market, and whether the design decisions actually support the business objective.
Over the last few years, however, I began to notice a pattern that became increasingly difficult to ignore. More and more websites started to feel strangely familiar. The logos were different. The industries were different. The colours had changed. Yet the structure was often almost interchangeable.
There would be a large hero section with an ambitious statement, followed by three or four rounded boxes. After that came a benefits section, then a feature section, then another row of cards, followed by perhaps a coloured gradient, some statistics, a frequently asked questions section and finally another call to action that repeated almost exactly what had already appeared at the beginning.
At first, I assumed that this was simply the result of contemporary design trends. Every era has its own visual language, and websites have always copied successful conventions from one another. But when I started reading the content carefully, another pattern emerged. The same point was often being made repeatedly in different forms. A promise introduced at the top appeared again in the benefits section, then once more in the feature section, then again near the bottom of the page.
The website looked polished, but it did not feel edited. It felt assembled.
That distinction matters.
A good website should not merely contain information. It should reflect decisions. Someone should have decided what matters most, what can be removed, what should be said only once, what deserves emphasis and what does not need to exist at all. Increasingly, I am seeing websites where those decisions appear not to have been made.
AI Is Very Good at Producing Completeness
One of artificial intelligence's greatest strengths is also one of its weaknesses. It is exceptionally good at producing something that feels complete.
Ask an AI system to create a website structure and it will rarely leave you with too little. It will give you a hero section, benefits, features, testimonials, process, use cases, FAQs and a closing call to action. Ask it to produce a proposal and it will supply context, objectives, methodology, milestones, deliverables and expected outcomes. Ask it to create a presentation and it will usually generate a neat sequence of problem, solution, framework, value proposition and next steps.
This completeness is useful. It saves time and reduces the friction of starting from nothing. But completeness can also create the illusion that the work is finished.
It is not.
The important work often begins after the generation.
An experienced human editor will remove repetition. A founder who understands the business deeply will recognise when a sentence sounds impressive but says very little. A designer with judgement will decide that six boxes should become three. A strategist will notice that a fashionable phrase is technically correct but inappropriate for the audience.
AI can provide material, but it does not automatically provide restraint.
The problem begins when people confuse generated completeness with finished quality.
What Worries Me Is the Surrender of Judgement
I am not opposed to artificial intelligence. Quite the opposite. I use it extensively, and I believe it is one of the most powerful productivity technologies available to us.
It can accelerate research, shorten development cycles, assist with coding, improve analysis, generate alternatives, organise information and help small teams perform work that previously required far greater resources. For entrepreneurs and SMEs in particular, this is transformative. A small business can now access capabilities that were once available mainly to large organisations with dedicated teams.
Yet there is an important difference between using AI as leverage and allowing AI to become a substitute for judgement.
That difference is becoming increasingly important because the modern workflow is dangerously easy. A business owner can ask AI to write a website, copy the result and publish it. A team can ask AI to prepare a proposal, change the company name and send it. A founder can ask AI to create a presentation, adjust the colours and present it to a customer.
Technically, the work is done. But intellectually, very little may have happened.
When this becomes normal, the human role gradually changes. Instead of thinking, shaping, challenging and deciding, the person becomes an operator who accepts or lightly edits what the machine has already proposed.
That is not efficiency. That is surrender.
Then I Started Hearing the Same Language Everywhere
The visual similarity of websites was only the first sign. The more interesting change appeared in business language.
Over the last few years, certain words have begun appearing with remarkable frequency. Operating System. Revenue OS. Business OS. Core. Centralized Control. Ecosystem. Framework. Engine. Unified Platform. Command Center.
None of these terms is inherently meaningless. Some are useful, and in the right context they describe real systems accurately. I have used some of them myself.
What becomes strange is the scale of repetition.
I have sat through presentations from different technology companies offering completely different products. One may be selling an AI platform, another a productivity tool, another an automation system, another a data solution. Yet the presentation structure often feels almost identical.
A central diagram appears. Something in the middle is labelled "Core". Several functions surround it. Another slide introduces the company's "Operating System". A later slide presents the "Ecosystem". Somewhere there is a "Framework", perhaps followed by an "Engine", a "Layer" or a "Centralized Control" concept.
After seeing this often enough, I began to experience something unexpected: instead of being impressed by the sophistication of the presentation, I became less interested.
The product might still be good. The company might be capable. But the language no longer tells me much about them.
I can see the vocabulary, but I cannot see the founder.
The More Intelligent Everyone Sounds, the Less Meaning Intelligence Carries
Artificial intelligence has made professional language dramatically more accessible. Someone who is not a natural writer can now produce elegant copy. Someone unfamiliar with presentation design can create polished slides. A small software company can describe itself using language that once belonged mainly to large consulting firms.
This democratisation is useful, but it also creates an unexpected consequence.
When everyone can sound strategic, strategic language becomes less distinctive.
When every proposal is polished, polish stops proving competence.
When every founder can produce sophisticated terminology, sophisticated terminology stops proving depth.
The result is that the market gradually begins to value something else: evidence of original thought.
A founder who has spent years solving a difficult problem tends to speak differently from someone who has only researched it. A builder who has implemented real systems usually talks differently about trade-offs, failure and limitations. Someone who has actually struggled through a business problem will often describe it with imperfect but highly specific language.
Experience leaves fingerprints.
Those fingerprints matter because they make a person credible.
Artificial intelligence, if used carelessly, can polish those fingerprints away.
AI Should Amplify Identity, Not Flatten It
I have long believed that AI amplifies what is already there.
When someone has strong experience, AI can help organise and extend that experience. When someone has good judgement, AI can help test alternatives. When someone has discipline, AI can make that person dramatically more productive. When someone is already a builder, AI can increase the speed at which ideas become systems, products and results.
But the reverse is also true.
If someone has no point of view, AI can amplify generic thinking. If someone does not understand the customer, AI can produce language that sounds persuasive but lacks insight. If someone has not developed judgement, AI can provide twenty alternatives without helping that person understand which one deserves to survive.
This is where the deeper risk of standardisation appears.
Artificial intelligence is trained on patterns, and therefore it is naturally good at producing plausible patterns. If millions of professionals rely heavily on those patterns without introducing enough of their own thinking, the output of the market will gradually converge.
More companies will become competent.
More websites will look professional.
More proposals will sound intelligent.
More presentations will appear strategic.
But more of them may also become interchangeable.
That is not progress if we lose identity along the way.
The Real Risk Is Not That Machines Become Human
Much of the public discussion about AI still revolves around machine consciousness. We imagine a future in which artificial intelligence becomes independent, develops intentions and eventually challenges human control.
That possibility receives attention because it is dramatic.
But a quieter and perhaps more immediate transformation is already happening.
AI does not need to become conscious if human beings become dependent.
There is a natural progression in how we use these tools. At first, we ask for assistance. We ask AI to help us write, summarise or improve something. Then we begin asking it what we should say. Later, we may ask what we should do. Eventually, without noticing the shift, we may begin asking what we should think.
That final transition matters enormously.
The distinctive human capacity is not merely the ability to produce text or create designs. Machines are already becoming very capable at those tasks. The more valuable capability is the ability to question assumptions, disagree with consensus, reject attractive nonsense and make decisions under uncertainty.
If we surrender that capacity, then AI does not need to defeat us.
We have voluntarily given away the most important part.
I Now Look for Evidence of Human Thinking
This has changed the way I evaluate work.
When I look at a presentation today, I am no longer impressed simply because it is polished. In fact, when the deck looks too familiar, I become more cautious.
I start asking different questions.
Does the language reflect something this company genuinely believes? Does the diagram actually describe how the product works? Can the founder explain why they chose this approach? Is there anything in the presentation that could only have come from their own experience? Is there a point of view here, or merely a collection of professionally arranged statements?
These questions matter because a presentation is not only communication. It is evidence of thinking.
When someone presents a company, I want to understand how they see the world. I want to know what they have noticed that others missed, what assumption they disagree with, what problem they understand unusually well and what difficult decision they have made.
I do not need every sentence to sound perfect.
In fact, sometimes imperfect language reveals more truth than polished language.
Human Judgement May Become the Scarce Resource
The cost of production will continue to fall. Websites will become cheaper to build. Videos will become easier to produce. Code will become faster to generate. Presentations, reports, marketing materials and research will become increasingly accessible.
As production becomes abundant, production itself becomes less valuable.
Judgement becomes scarce.
The person who can look at ten generated ideas and confidently reject nine will become valuable. The person who can distinguish between a clever sentence and a useful insight will become valuable. The person who can look at an attractive website and recognise that it does not solve the customer's problem will become valuable. The person who can remove half the content because the remaining half is stronger will become valuable.
Most importantly, the person who can tell AI, "This is not how I think," will become valuable.
That is why I do not believe the future belongs simply to people who know how to use AI. Eventually, almost everyone will know how to use it.
The advantage will belong to people who retain enough judgement to use it without becoming shaped by it.
We Should Build More With AI, Not Think Less Because of It
I want AI to help people build more ambitious businesses. I want it to help SMEs compete with much larger organisations. I want entrepreneurs to move faster, experiment more cheaply and create products that would previously have been beyond their resources.
That is the extraordinary promise of this technology.
But that promise depends on maintaining a distinction between assistance and authority.
AI should help us execute faster, but it should not become the owner of our convictions. It should help us explore ideas, but it should not become the source of our identity. It should help us see possibilities, but it should not remove our responsibility to choose among them.
This is why I continue to return to a simple principle in my own work: Dream It. Execute It. Ground It.
The dream must still come from a human understanding of what could exist. Execution can increasingly be accelerated by machines. But grounding — deciding what is useful, what is real, what is appropriate, what deserves trust and what should be rejected — remains a profoundly human responsibility.
The real threat of AI may therefore be much less dramatic than the stories we have imagined.
It may not arrive as a machine announcing that humanity is obsolete.
It may arrive quietly, through thousands of small decisions in which we stop questioning, stop editing, stop disagreeing and stop forming our own conclusions.
AI does not need to destroy humanity.
It only needs humanity to become comfortable enough to stop thinking for itself.