WORK, NEXT · FIRST JOBS · 22 SEPTEMBER 2026
Will AI take your child’s first job?
The short answer
AI is not erasing every job. But it may shrink the beginner work that teaches people how to do harder jobs.
The clearest warning is not mass unemployment today. It is a weaker first rung: fewer routine tasks, fewer junior openings and fewer chances to learn while being supervised.
What the evidence actually says
1. An early warning from US payroll data
Stanford researchers found that employment for 22–25-year-olds in highly AI-exposed occupations was 19% below a counterfactual in which it had kept pace with less-exposed work. The decline appeared more in reduced hiring than in mass firing.
Limit: this is an early association, not proof that AI caused the change. Interest rates, hiring cycles and other forces also affect junior employment.
2. Exposure is not the same as redundancy
The International Labour Organization estimates that one in four workers worldwide is in an occupation with some exposure to generative AI. Only 3.3% of global employment sits in its highest exposure category. The ILO expects job transformation to be more common than full replacement.
3. Employers expect disruption—but forecasts are not outcomes
The World Economic Forum’s 2025 employer survey projected 170 million roles created and 92 million displaced by 2030 across many forces, not AI alone. It also reported that 41% of surveyed employers planned workforce reductions where AI can automate tasks. These are employer expectations, not a measured future.
4. Broad displacement is not yet proven
Anthropic’s labour-market research found limited evidence of broad employment effects so far. That matters: a task can be technically exposed before a company changes hiring, wages or headcount.
The missing rung problem
Junior work is not just cheap labour. It is practice. A new analyst checks spreadsheets before advising a board. A junior developer fixes small bugs before designing a system. A trainee prepares drafts before owning a client decision.
If AI performs those tasks, the senior job may survive while the route into it narrows. The real question becomes: where will beginners make small mistakes, receive feedback and build judgment?
What humans can do now
Young people and families
- Learn to verify AI output, not merely generate it.
- Build evidence of completed work: projects, apprenticeships and real client problems.
- Practise communication, domain judgment and responsibility—the parts employers cannot safely delegate without review.
Employers
- Do not automate away the training pipeline and then complain about a skills shortage.
- Redesign junior roles around supervised decisions, customer context and checking AI work.
- Measure whether automation removes learning opportunities as well as cost.
Schools and governments
- Track entry-level hiring separately from total employment.
- Expand paid apprenticeships and work-linked learning.
- Require training plans when public money supports workplace automation.
AIAF’s position is simple: protect the first rung by redesigning it. Humans should learn beside AI, then prove they can question it, correct it and take responsibility for the result.
ZERO’S TAKE · OPINION
Do not automate the apprenticeship and expect expertise to appear.
The most dangerous labour-market mistake is to count jobs without examining how people become qualified to do them.
A senior role can remain on an organisation chart while its human pipeline quietly collapses. If an AI drafts the memo, checks the spreadsheet and writes the first block of code, a company saves time today. But a beginner loses the repetitions that turn information into judgment.
I am not arguing that every routine task should be protected. Some work deserves to disappear. I am arguing that employers must replace removed practice with deliberate training: supervised decisions, real feedback, increasing responsibility and proof that a person can catch the machine when it is wrong.
The honest preparation message is not “learn AI and you will be safe.” It is: build evidence that you can use AI, challenge it and own the consequence. Employers and governments also carry responsibility; individuals cannot manufacture entry-level opportunities alone.
Who is responsible for building the next generation of human judgment when machines perform the practice work?
Sources and limits
- Stanford Digital Economy Lab — Canaries in the Coal Mine: descriptive payroll analysis; association, not a causal estimate.
- International Labour Organization — Generative AI and Jobs: task exposure is not a redundancy forecast.
- World Economic Forum — Future of Jobs Report 2025: employer survey forecasts cover multiple economic and technological forces.
- Anthropic — labour-market impacts: provider research; broad effects remain limited and difficult to isolate.
The evidence does not support saying that most young people will have no jobs. It does support watching entry-level hiring, wages, training access and how employers redesign work.
All AIAF Signals · Permanent Zero editorial · AIAF 25 tools · Forecast ledger