The daughter of a Microsoft AI tech bloke was choosing careers, so he ranked them.
First problem: how would she even get started?
The most important thing to understand [he discovered] is also the most counterintuitive. AI is not demolishing careers from the top. It is removing the bottom rungs of the ladder first.
Think about how almost every professional career develops. A “Big Four” trainee at the accountancy firms Deloitte, KPMG, EY or PwC reconciles spreadsheets and drafts standard documents. A junior solicitor reviews contracts. A graduate analyst builds financial models. These are the apprenticeship stages. They are how young professionals develop the judgment that eventually makes them irreplaceable.
AI performs many of those tasks faster and more cheaply.
Robots are going to take your job? No doubt.
What if robots take all the jobs? Hint: They can't.
Comparative advantage tells us that "new kinds of jobs will appear, as they always have when technology advances."
Ironically, most of the jobs people are afraid of losing -- such as programming jobs or truck-driving jobs -- were themselves created by technological advances....
What new types of job will be created? I can no more project that than a man in 1956 could have projected that today there would be jobs in something called “social media”; or that money can be made by driving for Uber and by renting out living space through AirBnB.
One estimate is that alongside all the jobs displaced, around 170 million will be created. "The net picture is not collapse. It is transformation. And transformations reward the families who understand them early."
What's to understand, says the tech bloke, is the four things in which human beings do have a comparative advantage over any machine: Emotional intelligence. Creative vision. Physical dexterity. Ethical judgment. Based on that insight, the tech bloke ranked careers
across nine categories including emotional intelligence, creative thinking and vulnerability to AI tools. A score closer to 100 per cent means the role depends heavily on things AI cannot replicate. A score closer to 35 per cent means much of the work is already within reach of automation.
Biggest winners:
- healthcare
- education
- skilled trades
- creative industries (for genuine creatives)
- tech, finance and law (for those at the level required to exercise judgement)
- diplomacy
- paralegal
- accountant
- data entry
- admin
We are facing a particular moment in history. It is not one that will announce itself. There is no letter from school, no official notification that the world your child is preparing for has quietly become a different one.
The families who will look back on this decade without regret are the ones who had the conversation early and trusted that a child who understands the world they are entering is far better equipped than one protected from it. Here are some first steps:
If your child is 10 to 12, build the foundations: teach them to be curious by reading carefully and arguing a point. Curiosity is the hardest quality to automate.
If your child is 13 to 15, have one conversation this week. Not a lecture. Ask what they think AI is doing to the world. Help them begin using AI tools, not to do homework for them, but to understand what these systems cannot do. That understanding is the first superpower.
If your child is 16 to 18 and making real choices, look hard at where the four human superpowers appear in the careers they are considering. AI fluency is not optional any more. The wage premium for those who have it is visible and growing fast. “Wait and see” is not a neutral position. It is a decision. The data says it is the wrong one.
NB: The Economist magazine's analysis suggests AI may already be harming some graduates’ job prospects
We found that graduates in fields more exposed to AI have suffered markedly worse outcomes. Between 2022 and 2024 graduates in the least-exposed quintile—studying subjects such as education, philosophy and civil engineering—saw their average full-time employment rate fall by just 1.5 percentage points. Those in the most exposed quintile—including computer science, computer engineering and information science—suffered a 6.6 percentage-point drop (see chart 1 above). ... the trend continued for the class of 2025 (see chart 2 below).
"Which jobs can AI learn to do? We examine this for every occupation in the US economy."What Jobs Can AI Learn? Measuring Exposure by Reinforcement Learning - CORNELL UNIVERSITY RESEARCH PAPER, 4 May 2026
"We investigate the potential implications of large language models (LLMs) on the U.S. labour market."An Early Look at the Labor Market Impact Potential of Large LanguageModels - UNI OF PENNSYLVANIA, Aug 2023






.jpg)
















