The working-age population of developing economies in Europe and Central Asia could shrink by around 17 million by 2050. Artificial intelligence could partly offset that decline through productivity gains and task automation. The benefits will depend on employment structures, workers’ skills and how the technology is introduced, while shortages in some occupations may coexist with job cuts in others.
Ageing populations make productivity an increasingly important source of economic growth. With fewer potential workers, sustaining output and services requires more efficient use of labour. AI can help by taking over some duties and reducing the time needed for others.
World Bank economists Ivailo V. Izvorski and Filip Jolevski discuss these opportunities in an analysis published on VoxEU on 30 October 2026. They argue that AI could cushion the economic effects of demographic change. The scale of the gains will also depend on investment in skills and changes in how work is organised.
A shrinking labour supply increases pressure to improve productivity
Projections cited in the analysis suggest that by 2050 there will be around 40 older people for every 100 working-age people in the region, almost twice the current ratio. The median age is expected to rise from 36 in 2023 to around 49 in 2060. These projections cover the developing economies of Europe and Central Asia included in the World Bank analysis, rather than Poland alone or the entire European Union.
A smaller workforce can constrain companies’ expansion even when demand for their products holds up. Automation can reduce the labour required for particular tasks. It may also help businesses keep operating after some staff retire.
Higher labour force participation also matters. Women’s participation remains low in some countries in the region, and older workers leave employment earlier than in advanced economies. Making greater use of this potential workforce depends on childcare, health, employment conditions and training; AI can support longer working lives but does not remove those constraints.
Fewer jobs do not eliminate labour shortages
A long-term decline in labour supply does not rule out temporary increases in unemployment. Employment also depends on the business cycle, operating costs and changes in sectoral demand. Mismatches in skills and location can leave vacancies unfilled even as the number of jobseekers rises.
According to Statistics Poland (GUS), average enterprise-sector employment in September 2026 was 6,360,400 full-time equivalents, down 0.8% year on year. The decline over the first nine months was 1%. At the end of September, 907,000 people were registered as unemployed, 4.7% more than a year earlier, and the registered unemployment rate was 5.8%.
Sectoral trends differed. Employment in January–September fell by 1.6% in manufacturing and 1.4% in construction, but rose by 2.1% in accommodation and food services. Enterprise-sector statistics cover only part of the economy and cannot isolate AI’s effects from other influences on labour demand.
AI can replace tasks and complement workers’ skills
AI’s impact on employment depends on whether it takes over workers’ duties or mainly helps them perform those duties. In the first case, a business may need less labour to produce the same result. In the second, employees may complete existing tasks faster or take on more complex work.
A study cited by the World Bank economists estimates automation potential at around 11% of jobs in Bulgaria, Croatia, Poland and Romania. Examples include document processing, standard customer service, basic accounting and some programming tasks. This measures technological potential rather than predicting job losses.
Separate estimates cited in the analysis put around 27% of EU jobs in a category where AI can primarily complement human work. The authors identify a similar proportion in Europe and Central Asia. Judgement, understanding context and responsibility for decisions remain important in these roles.
The employment outcome also depends on how demand responds. Lower service costs may attract more customers and allow expansion; if demand is unchanged, productivity gains may instead reduce hiring or headcount. Businesses with unfilled positions may use those gains to ease staff shortages.
Highly skilled occupations are also exposed
The potential changes are substantial in Poland. A Polish Economic Institute (PIE) study published in 2024, using 2022 Labour Force Survey data, found that around 3.68 million people worked in the 20 occupational groups most exposed to AI. They represented 22% of the workers covered by the analysis, and 82% had higher education.
Exposure was high among finance professionals, lawyers, programmers and administrative specialists. The indicator compares the scope for using AI across occupations. It neither identifies jobs earmarked for elimination nor determines whether AI will replace workers or improve their productivity.
For labour shortages, the relevant question is which duties can be automated. More efficient administration can free up people providing services, but cannot replace every task requiring physical presence, direct personal care or manual work. Productivity gains in one part of the economy therefore do not automatically resolve shortages elsewhere.
Slower hiring may precede a fall in employment
AI’s labour-market effects may initially appear in recruitment. A study described on 8 October 2026 by Simone Maria Bonin, Saman Darougheh and Andreas Kuchler linked data on Danish companies’ AI use with monthly employment registers.
Companies that began using AI in 2023 subsequently expanded employment more slowly relative to their earlier trend than businesses that did not use the technology. The adjustment mainly involved fewer hires, particularly in roles highly exposed to AI. The effect was statistically significant in smaller firms, while employees with longer tenure were more likely to stay.
The findings suggest entry into the labour market could become harder even without a clear increase in dismissals. They do not, however, establish causation: companies chose when to adopt AI, and technology and staffing decisions may have been made together. Results from Denmark cannot be directly transferred to Poland.
Experience can complement the technology
Automating tasks increases the importance of skills needed to assess the results. Knowledge of an industry, its customers and unusual situations can help workers spot errors and determine whether a system’s answer addresses the actual problem. Experienced employees may therefore apply their expertise within a different division of duties.
Age or tenure alone does not determine the benefits of AI. Tasks, digital skills and access to training also matter. Excluding older employees from skills development can limit the use of their knowledge and make it harder for them to remain in employment.
Flexible hours, part-time work and gradual handovers can help people approaching retirement stay active. Their health and family circumstances also affect what is possible. Longer working lives require changes in employment conditions as well as access to new tools.
Economic gains depend on adoption and implementation
Simulations presented by the World Bank economists suggest AI’s impact could grow as adoption spreads. In an optimistic scenario, regional growth in 2027 would be around 0.1 percentage points higher than without AI. The additional contribution could reach around 2 percentage points annually in 2033–2034; these are regional model scenarios rather than forecasts for Poland alone.
Technological potential does not guarantee adoption. Data cited in the analysis show that fewer than one in ten businesses in the economies studied use AI, around half the EU proportion. Interviews with managers in five countries during summer 2026 found that companies mainly focused on improving internal processes and cutting costs, using AI less often to expand their offerings and increase revenue.
Telecommunications and finance lead adoption, while retail and less technologically advanced manufacturing businesses remain more cautious. Barriers include insufficient skills, poorly organised data and difficulty estimating returns on investment. Speeding up one task does not necessarily improve an entire process, and checking and correcting results can absorb some savings, making quality and supervision costs important.
Changing duties also requires different training. Employees need both practical tool skills and the knowledge to assess outputs, while those whose tasks are taken over may need preparation for other roles. Removing simple entry-level duties creates a further challenge: companies must still let new employees learn through real work under experienced colleagues’ supervision.
Polish original published on 30 October 2026: CEO.
