AI can automate tasks, but human skills, presence and judgement remain difficult to replace.
When we talk about Artificial Intelligence, the first question that usually comes to mind is simple:
“How many jobs will AI take away?”
I think we may be looking at the question from the wrong end.
For me, the more interesting question is:
“Which skills can AI replace, and which skills still need a human being?”
That is why the Iceberg Index caught my attention.
The research behind the Iceberg Index looks at the U.S. workforce through the skills involved in different occupations. Its important finding is that visible AI adoption is only a small part of the story. The study estimates that AI’s technical capability already overlaps with skills representing about 11.7% of U.S. labour-market wage value, roughly $1.2 trillion. But this is a measure of technical exposure. It is not a prediction that 11.7% of workers will lose their jobs.
And this is where the iceberg becomes interesting.
The AI pyramid I see today
If I were to describe today’s workforce as a simple pyramid, I would put it this way.
At the very top — 25% — are the high or fully involved AI explorers.
These are people whose work is already deeply connected with AI, digital tools, data and automation.
In the middle — 45% — are intermediate explorers.
They are using AI for some tasks. They may use it for research, writing, analysis, communication or routine office work. But they still depend heavily on their own experience and judgement.
At the broad base — 30% — are people with little or no exploration of AI so far.
Think about a cook, mechanic, plumber, bartender, nurse, childcare worker or many other hands-on occupations.
This pyramid is not a percentage breakdown published by the Iceberg Index. I see it as a simple way of understanding how unevenly AI is entering working life.
And that difference matters.
AI can change a job without replacing the person
A mechanic may use AI to identify a fault.
A nurse may use digital tools to organise patient information.
A cook may use AI to develop a recipe.
A plumber may use technology to diagnose a problem.
But the physical work, personal judgement, human contact and responsibility still remain.
That is why I do not see AI simply as “humans versus machines.”
I see it more as AI taking over certain skills and tasks while humans continue to provide other skills that machines cannot easily deliver in the real world.
This is perhaps the most important message hidden beneath the iceberg.
The future may favour skill-centred work
There is a lesson here that I find particularly interesting.
The more a job depends on information that can be processed digitally, the more AI can potentially enter that job.
But when work depends on hands, presence, judgement, trust, physical movement and personal interaction, the situation becomes different.
This does not mean such jobs are completely protected from AI.
It means the path to automation is harder.
And that may give skilled workers in these occupations continuing value.
A good electrician, experienced mechanic, skilled barber, competent nurse or trusted childcare worker is selling something more than information.
They are selling human skill and human presence.
Then comes Baumol’s Cost Disease
This brings me to another fascinating economic idea: Baumol’s Cost Disease.
The basic idea is quite easy to understand.
Some industries become dramatically more productive because technology allows one worker to produce much more.
Manufacturing is a classic example.
But a teacher still has to teach. A nurse still has to care for patients. A childcare worker still has to look after children. A barber still has to cut hair.
Their productivity cannot always increase at the same speed as that of a highly automated factory.
Yet their wages still have to compete with wages elsewhere.
Economists have studied this phenomenon for decades. Research has also found evidence of Baumol-type effects in healthcare, although economists differ on how much of rising healthcare costs can actually be explained by it.
Why some things become cheaper while others become expensive
Look around us.
Computers have become enormously more powerful.
Digital storage has become cheaper.
Communication has become faster.
Many digital services can now be delivered almost instantly.
Technology increases productivity in these areas.
But healthcare, education, childcare and many personal services still require large amounts of human time.
That is why the economics can move in the opposite direction.
The things that technology can produce more efficiently tend to become cheaper or more productive.
The things that still depend heavily on human labour can remain expensive.
That is the fascinating connection between AI and Baumol’s Cost Disease.
My takeaway
I don’t think the future is simply about AI replacing jobs.
The bigger story is about AI changing the value of different skills.
Some routine skills will become less valuable because machines can perform them faster.
Other skills may become more valuable precisely because machines cannot easily reproduce the human element.
After 38 years of watching people work, communicate, sell, negotiate and build relationships, I have seen one thing repeatedly:
A technology may change the way we do a job. But the value of a good skill does not disappear simply because technology arrives.
The Iceberg Index therefore gives me a different way of looking at the AI revolution.
The visible part is AI.
The much bigger question underneath is human skill.
And perhaps that is where the real future of work will be decided.
#IcebergIndex #AIandJobs #FutureOfWork #HumanSkills #ArtificialIntelligence