Publication date: 12 October 2026
Artificial intelligence may be creating a specific labour-market problem before it produces mass unemployment: weaker demand for people trying to enter professional careers. Singapore President Tharman Shanmugaratnam told the IMF and World Bank annual meetings that the first job after graduation is already becoming harder to secure in sectors most exposed to AI.
The warning matters because entry-level work is not only a source of income. It is where graduates learn how organisations operate, receive supervision, build professional judgement and become candidates for more senior roles. If AI automates the routine tasks traditionally assigned to junior employees, companies may save money today while weakening the pipeline of experienced workers they will need later.
The International Monetary Fund is presenting research based on millions of online job postings across Asia to examine how AI is reshaping labour demand. IMF Managing Director Kristalina Georgieva has said that close to 40 per cent of jobs globally are exposed to AI-driven change, with a higher share in advanced economies.
Automation starts with tasks, not whole occupations
The immediate effect of AI is unlikely to be the disappearance of entire professions. It is more often the automation of individual tasks: drafting routine documents, summarising information, writing basic code, conducting initial research or responding to standard customer requests.
Those are precisely the activities that often help new employees learn. Removing them can raise the productivity of experienced staff, but it can also make junior roles harder to justify. Employers may recruit fewer graduates and expect those they do hire to contribute at a higher level from the first day.
This produces a paradox. Businesses say they need workers with practical AI skills, yet fewer entry-level opportunities mean fewer places where young people can acquire those skills in a real organisational setting. Universities can teach tools and concepts, but they cannot fully replicate responsibility, teamwork, client pressure and the consequences of operational mistakes.
The risk of a divided labour market
The disruption will not be evenly distributed. Advanced economies and large companies are better placed to invest in AI and redesign jobs around it. Smaller firms and developing economies may adopt more slowly, but they may also struggle to capture productivity gains. That can widen differences between countries, businesses and workers.
Within companies, the greatest rewards may flow to experienced employees who use AI to amplify existing expertise. New entrants without professional context may find that access to the same tool does not produce the same value. Knowledge of what to ask, how to verify an answer and when to reject it comes from experience that technology cannot instantly supply.
The consequences extend beyond graduate unemployment. If fewer young people enter skilled occupations, tax revenues may weaken, wage progression may slow and social mobility may decline. Firms could later face shortages of mid-career talent because they failed to train enough people at the beginning of the pipeline.
Redesigning the first job
The answer is not to preserve repetitive work for its own sake. Employers need to redesign entry-level roles so that AI supports accelerated learning rather than replacing the learner. Junior staff can use automated tools while being assessed on verification, judgement, communication and domain understanding.
Structured apprenticeships, rotations and supervised project work will become more important. Organisations should also track hiring and promotion by career stage, not only total headcount. Stable employment numbers can conceal a collapse in entry routes if experienced hiring replaces graduate recruitment.
Governments and universities need shorter feedback loops between job-market evidence and course design. Technical literacy should be combined with sector knowledge, critical evaluation and interpersonal skills. The objective is not to train every graduate as an AI engineer, but to ensure they can work responsibly in organisations where AI is embedded in everyday decisions.
The first rung of a career ladder has always been less productive than the levels above it. Its value lies in creating future capability. Leaders who remove it without building a replacement may discover that automation has solved a short-term cost problem by creating a long-term talent problem.