AI Was Supposed to Make Us Smarter. What If It’s Doing the Opposite – and Fast?
I am a proponent of AI.
I use it. I encourage clients to use it. I think it is one of the most consequential technologies we have seen in decades, and I have no interest in pretending we can, or should, put the genie back in the bottle.
But something has been happening in my work recently that has genuinely stunned me.
I work closely with a number of clients and their marketing teams, and I am seeing a pattern emerging. Not occasionally. Increasingly, strongly and alarmingly.
People are losing confidence in their own ability to think.
They are outsourcing, not just the execution of work to AI, but the thinking that is supposed to happen before the work.
And it has happened remarkably quickly.
The Illusion of Good Work
There is an obvious efficiency argument for AI.
Why spend three hours writing something when AI can produce a first draft in seconds?
Why manually research something when an AI tool can synthesise information almost instantly?
Why stare at a blank page when you can ask a machine to give you ten ideas?
All of this makes sense.
The problem starts when the AI stops being a tool for thinking and becomes a substitute for thinking.
I am seeing people increasingly ask AI what they should think rather than using it to help them think.
They ask it to solve the problem before they have properly defined the problem.
They ask it to develop the strategy before they have developed a point of view.
They ask it to write the answer before they have worked out what they actually want to say.
And then, perhaps most concerningly, they accept the answer.
There is no interrogation.
No disagreement.
No refinement.
No moment of: Hang on. That doesn’t feel right.
The machine has produced something plausible, so the work feels finished.
But plausible is not the same thing as good.
We are Outsourcing Judgement
This is where I think the real danger lies.
The most valuable part of professional work has never been the ability to produce words, slides, spreadsheets or documents.
It is judgement. Knowing what matters. Knowing what doesn’t.
Recognising when something is technically correct but strategically wrong.
Understanding the context that isn’t written down.
Seeing the contradiction.
Spotting the opportunity.
Knowing when to challenge the brief.
Knowing when an idea is interesting but won’t work.
Knowing when something sounds impressive but actually says nothing.
AI can help enormously with all of these things.
But it cannot relieve humans of responsibility for them.
When we continually outsource these decisions, something else happens: our own judgement begins to weaken.
Use It or Lose It
Think about any skill.
If you stop practising it, you get worse.
If you always use a calculator, you become less comfortable doing mental arithmetic.
If you never navigate without GPS, your sense of direction can deteriorate.
If you never write without autocomplete, you may become less confident constructing a sentence yourself.
If you never pick up a pencil, you may never be capable of hand – writing anything – something already evident in the younger generations.
Thinking is no different.
If you continually hand the difficult cognitive work to a machine, eventually you become less accustomed to doing that work yourself.
And perhaps less confident that you can.
What Happens in the Absence of Human Supervision
Once people become accustomed to AI doing the thinking, they also become less capable of reviewing the output.
This creates a particularly dangerous loop.
AI produces something. The human reviews it.
But the human no longer has enough confidence, knowledge or critical distance to properly evaluate it.
So the AI’s output becomes the benchmark against which itself is judged.
That is not oversight – it is automation without meaningful supervision.
And it is particularly easy to miss because the output often looks good.
The grammar is correct.
The formatting is polished.
The argument appears coherent.
The tone sounds professional.
There are no obvious errors.
But underneath the surface, there may be nothing particularly original, insightful or useful.
The work is fine.
And “fine” is becoming one of the biggest risks of the AI era.
Marketing is Especially Vulnerable
There is a seductive idea spreading through business that because AI can write, companies no longer need copywriters, communicators or marketing specialists.
I think this fundamentally misunderstands what those people are actually paid to do.
A great copywriter isn’t simply someone who can assemble grammatically correct sentences.
A great marketer isn’t someone who can produce a thousand social posts.
A communications professional isn’t simply a human interface for a language model.
The craft is much deeper than that.
Marketing sits at the intersection of strategy, psychology, culture, language, behaviour and business.
It requires an understanding of what people say they want versus what they actually respond to.
It requires nuance.
It requires context.
It requires taste.
It requires knowing when to say something and, equally importantly, when not to say it.
It requires an instinct for the difference between an idea that is merely clever and one that will actually change someone’s behaviour.
It requires understanding the customer as a human being rather than as a data point or a prompt.
AI can support all of this.
But it doesn’t magically acquire the experience, judgement and instinct of the people who have spent years developing those capabilities.
AI Doesn’t Replace Expertise. It Amplifies It.
This is the part I think businesses are getting wrong.
The arrival of AI doesn’t necessarily make expertise less valuable.
It can make expertise more valuable — but only when expertise remains in the loop.
If AI allows one experienced marketer to do the work that previously required five people, that doesn’t mean the answer is necessarily to remove the experienced marketer.
It means that person’s judgement can now be multiplied.
One person can explore more ideas.
Test more variations.
Analyse more information.
Move more quickly.
Produce more iterations.
Develop more sophisticated campaigns.
AI can dramatically increase the leverage of expertise.
But there needs to be expertise there to leverage.
Without it, you simply get more output.
And more output is not necessarily more value.
An Ocean of Content
There is another consequence that concerns me.
We are about to be flooded by an extraordinary amount of content that nobody particularly needs.
AI has made the marginal cost of producing content approach zero.
So organisations are going to produce more of it.
More articles.
More emails nobody asked for.
More social posts nobody reads.
More thought leadership that says nothing new.
More white papers nobody downloads.
More “insights” that contain no insight.
More content that sounds authoritative while adding absolutely nothing to the body of
knowledge or the conversation.
When Content Becomes Cheap, Ideas Become Scarce
An enormous proportion of it will be technically competent and intellectually empty.
AI slop.
And the irony is that businesses pursuing efficiency may end up making themselves less effective
Because when everyone can produce content, content itself becomes cheap.
What becomes scarce is something else.
A genuinely interesting idea.
A distinctive point of view.
A useful insight.
A human observation.
A piece of writing that understands its audience.
A campaign that makes someone feel something.
A strategy based on actual understanding rather than generated consensus.
In other words: original thinking.
The Competitive Advantage Won’t Be AI
Eventually, almost every serious organisation will have access to broadly similar AI capabilities. The tools will become cheaper, the models more capable and the interfaces easier to use. The competitive gap created simply by having AI will disappear.
So where will competitive advantage come from?
I think it will come from what organisations do with it.
The businesses that win will be the ones that combine powerful AI with exceptionally skilled humans.
They will use AI aggressively — but intelligently.
They will automate the repetitive.
Accelerate the research.
Generate possibilities.
Challenge assumptions.
Analyse information.
Create iterations.
But they will keep humans responsible for judgement.
They will retain people who understand customers, markets, psychology, language, culture and strategy.
People who can look at an AI – generated answer and say: No.
That’s technically good, but it’s wrong.
That’s interesting, but it won’t work.
That’s what everyone else is saying.
That’s not what our customers actually care about.
There’s a better idea here.
That human ability to discriminate between good and bad, useful and useless, original and derivative, persuasive and merely polished – that is going to become incredibly valuable.
Don’t Let the Machine Become the Thinker
I don’t think the answer is to use AI less. The answer is to use it better.
Use it to challenge your thinking, not replace it.
Ask it for alternatives after you have formed your own view.
Use it to find weaknesses in an argument you have developed.
Ask it to identify things you may have missed.
Use it to accelerate execution once the strategic thinking is done.
Let it help you explore.
Let it help you create.
Let it help you work faster.
But don’t hand over your judgement.
Don’t let the convenience of an instant answer eliminate the uncomfortable process of working out what you actually think.
And don’t confuse a machine’s ability to produce an answer with your responsibility to decide whether that answer is worth anything.
Because that distinction is going to matter enormously.
We are entering a world where producing content will be incredibly easy.
Thinking will not.
Critical Thinking Is the Competitive Advantage
I suspect that, before long, the organisations that have protected and developed their people’s ability to think critically will have a profound advantage over those that have simply automated it away.
The future doesn’t belong to businesses that use AI instead of skilled people.
It belongs to businesses that use AI through skilled people.
The former will produce more.
The latter will produce what matters.
And while everyone else is drowning in AI slop, the organisations that still know how to think will be charging ahead.