The training pipeline that turned graduates into working professionals is breaking, and no one agrees on who should repair it. A June 2026 BBC News panel, “Is Gen Z trapped by AI?”, put the problem plainly. Young people were told AI would make them more employable, and instead many watch the first rung of the ladder disappear before they can reach it. The question the panel opened is the one worth sitting with. If entry-level work is going, who builds the replacement, employers, schools, or governments.
The scale is global, not a Western anxiety. The ILO’s Global Employment Trends for Youth 2026 puts the youth unemployment rate at 12.4 percent and the share of young people not in employment, education or training at 20 percent, or roughly 257 million people. The report names the exact roles that used to be first jobs, clerical, administrative, sales and junior technical positions, as the ones now contracting fastest.
What the AI numbers actually show
The strongest evidence is narrower than the headlines suggest. Stanford’s Digital Economy Lab, using US payroll data, found that workers aged 22 to 25 in the most AI-exposed occupations are now employed about 19 percent below where the trend would place them, while experienced workers in the same fields show no comparable gap. The mechanism matters. It runs through reduced hiring of young people, not through firing the ones already employed.
That is the tension the BBC panel circled. Sarah O’Connor, author of “We Are Not Machines,” and Rumman Chowdhury of Humane Intelligence both pushed at the same worry, that the routine tasks AI now absorbs are precisely the tasks juniors once did to earn their place. The result is a hiring freeze aimed at one end of the workforce.
The pipeline problem employers built
Firms cut junior roles because software now handles the low-stakes work that used to justify them. Those same junior roles were also how companies grew their own senior staff. Remove hiring and training at the bottom, and the supply of experienced workers thins out a decade later. This is the training question underneath the unemployment numbers, and it lands on employers first because they own the jobs.
Who has actually built a bridge
A few systems already answer the question, and they answer it differently. Germany routes roughly half of its school leavers through a dual vocational training system that splits time between a company and a classroom, and its youth unemployment rate sits near 5.7 percent, the lowest in Europe. Here the employer carries the training, with cost shared across firms and the state.
Elsewhere the state moves first. JobBridges recently covered West Africa’s publicly funded school-to-work programs, built precisely because private employers were not hiring or training at the scale the youth population demanded. Two models, employer-led and state-funded, and then the arrangement most of the world actually has, which is neither.
The argument, without a verdict
Each side has a real case. Employers created the gap by cutting the roles, so the responsibility to rebuild training sits with them. Education systems counter that they cannot teach job-specific skills for jobs that shift every two years, and that firms are the only ones close enough to the work. Governments face the bill either way, through youth unemployment support or through subsidised apprenticeships, which is the argument for acting early rather than paying later.
There is no settled answer, and the panel did not reach one. What the evidence supports is narrower and harder. The gap is real, it is concentrated at the entry point, and the countries that have closed it did so by deciding, explicitly, who pays to train a beginner.
Germany trains close to half its school leavers through companies. Most countries train almost none, and the BBC panel, for all its expertise, could not name who changes that.