Entry-level jobs for young workers in AI-exposed occupations

Brynjolfsson, Chandar and Chen of Stanford use payroll data from ADP to follow millions of US workers, up to and including June 2026. They first documented the effect in August 2025 and revised the paper in August 2026. They find no evidence of broad job losses due to AI. But employment of 22 to 25-year-olds in occupations highly exposed to AI is 19 per cent lower than if it had kept pace with peers in less exposed occupations, and that gap has widened steadily since August 2025. Older workers in the same occupations show no comparable gap. The effect runs through less hiring; layoffs are not rising. It is concentrated in occupations where AI replaces human tasks; where AI complements the work, employment is flat or rising. The adjustment is in the number of jobs; base pay does not move with it. The authors call these descriptive signals rather than causal estimates: the pattern holds when tech companies are excluded, weakens partly after controlling for education and is stronger in the ADP data than in national surveys.

For its AI Jobs Barometer 2026, PwC examined more than a billion job ads in 27 countries and, for the first time, looked separately at entry-level roles, using 2.4 million US vacancies. The entry-level roles most exposed to AI are seven times as likely to ask for skills that used to belong to senior staff: judgement, leadership, creativity, dealing with people. These roles have grown by 35 per cent since 2019, while other entry-level roles shrank by 10 per cent. The bottom rung, in other words, is getting higher. More broadly, PwC sees a two-track labour market: roles in which AI augments the expert are growing twice as fast as roles that AI opens up to non-experts. The wage premium for AI skills rose to 62 per cent, and the companies best able to use AI are growing headcount faster than the least exposed (52 against 36 per cent).

Deloitte asked the new starters themselves, surveying 22,595 young people in 44 countries. Three quarters use AI daily at work. One in five Gen Z respondents say their organisation is cutting entry-level roles, almost a quarter see new ones emerging, and the effects mentioned most often are gaining experience faster and having more time for work that matters. Almost a third think their organisation is not prepared for what AI will bring. McKinsey, meanwhile, hears from 39 per cent of organisations that they expect AI-related cuts to their workforce in the coming year, against 32 per cent a year earlier, while only half of the cuts expected the year before actually materialised.

In January 2026 Lynda Gratton of London Business School puts the underlying question on the table. In her view, AI also changes the route by which people build expertise. The repetition, the struggle and the refining through which managers themselves became skilled professionals threaten to disappear from the work. She asks leaders to consider what happens if those paths to mastery fall away. She writes about employees in general and does not single out new starters.

For the European car industry, ACEA and Adecco expect a shift towards engineers, IT staff and managers by 2035, and less demand for intermediate and lower-level work. This comes from their analysis of 39 automotive regions, published in January 2026. The group of associate professionals, which includes entry-level roles in administration, IT and engineering, is projected to shrink in Styria and in Czech and French regions and to grow in Italy and Spain. The report looks at the whole workforce and does not break it down by age.

How solid is this? Three independent institutes find the same pattern in 2026 from three angles (payroll data, vacancies, the new starters themselves), and with a year of monthly figures Stanford has the first real series. The cause is still disputed and Dutch figures are missing, so the picture is not settled.