Artificial Intelligence and the Future of Entry-Level Work
A · Strong evidence A World Economic Forum and PwC report combining ILO employment data, PwC's job-postings barometer and a survey of more than 9,000 entry-level workers in 48 countries, with employer case studies.
Summary based on the full text·Original: about 50 minutes to read
In short
A World Economic Forum insight report produced with PwC (June 2026) on how AI changes entry-level work. It finds entry-level hiring is slowing, most sharply in AI-exposed occupations, though declines in postings began before ChatGPT and employers often use AI as the stated reason for cuts driven by cost and uncertainty. The report sets out a four-part framework (job access, job design, talent pipelines, education alignment). It asks employers to treat entry-level hiring as an explicit part of workforce planning rather than a line to cut.
Key numbers
- 37%young workers (15-24) globally in occupations with medium to high exposure to AI-driven task change
- 16%decline in U.S. entry-level jobs in AI-exposed fields since late 2022 (Brynjolfsson et al., cited)
- 68% / 45%entry-level workers reporting higher productivity from AI / spending more time working because of AI
- 28%entry-level workers who expect half or fewer of their skills to stay relevant in three years
As reported by the source.
English·PDF·opens reports.weforum.org
What it reports
- 37% of young workers worldwide are in medium or high AI-exposure occupations: 75% in Eastern Asia, 69% in Northern America, 63% in Europe. Finance, information and communication, and professional services are the most exposed sectors.
- Brynjolfsson et al. find a 16% decline in U.S. entry-level jobs in AI-exposed fields since late 2022. PwC data show early-career postings in the highest-exposure quartile flatlining, with declines that started nearly a year before ChatGPT.
- 68% of entry-level workers report higher productivity from AI, but 45% also work more hours as a result.
- Companies with the strongest AI-driven financial outcomes are twice as likely to redesign workflows.
- 23% of CEOs say current structures constrain performance. Expected AI-related restructuring at entry level is almost twice that for mid or senior roles, and three-quarters of leaders expect significant realignment.
- Entry-level roles in the top AI-exposure quartile show nearly twice the net skills change of non-entry roles. 28% of entry-level workers think half or fewer of their skills will be relevant in three years, and 31% plan to ask for a promotion next year.
- More than 20% of entry-level workers are changing careers, re-entering work or moving industries later in life.
- Dropbox expanded intern and new-graduate programmes by 25%; 65% of those hires were promoted by year two and 87% by year three.
- In the U.S., 42% of Gen Z workers are in or pursuing blue-collar or skilled-trade roles, including 37% of those with a bachelor's degree.
How they got it
The report draws on the Forum's Global Dialogue on AI and Entry-Level Work (more than 200 leaders and experts), PwC's 2026 AI Jobs Barometer (occupation exposure scores matched to ILOSTAT youth employment), PwC's Global Workforce Hopes and Fears Survey 2025 (more than 9,000 entry-level workers in 48 countries), employer interviews and working sessions.
Limits
- "Exposure" measures task overlap with AI capabilities, not actual displacement.
- The worker data is self-reported perception; company cases such as Dropbox's promotion rates are the firms' own figures.
- PwC sells workforce consulting, and the barometer is its own product.
Why it matters for companies
Cutting graduate intake to bank AI savings removes the pipeline that produces tomorrow's managers and experts. The report's practical asks are concrete: set intake targets in workforce plans, decide which tasks stay with juniors because they build judgment, and define progression for flatter structures. The 45% working longer is a warning that "productivity" can mean extra load instead of better work.
Summary based on the full text. Read the original for the full argument and data.Open PDF on reports.weforum.org