The World Economic Outlook, published by the International Monetary Fund, highlights that artificial intelligence (AI) is reshaping the global economic landscape at an accelerating pace (IMF, 2025). The adoption of AI is generating significant macroeconomic spillovers, including shifts in productivity within AI-producing sectors and major changes to labour market structures and energy demand. This trend is already evident in countries such as the USA, where the value added by AI-producing sectors increased from $278 billion in 2010 to $1.13 trillion in 2023, quadrupling in just over a decade and raising their share of GDP from 2.4% to 3.5%. This growth rate surpasses that of traditional manufacturing and the private non-farm sectors.
The diffusion of AI is generating profound transformations in labour markets, with heterogeneous effects across occupations, age groups and education levels. As the IMF report shows, the impact of AI depends largely on whether it substitutes or complements human labour. The framework distinguishes three types of job: high exposure and low complementarity (HELC); high exposure and high complementarity (HEHC); and low exposure (LE). Jobs in the HELC category are more susceptible to displacement driven by automation, while those in HEHC occupations benefit from AI augmentation and tend to experience increases in productivity and wages.
An analysis based on microdata from the USA and Brazil[1] finds that around 45% and 65% of workers, respectively, are employed in high-end cognitive (HEHC) occupations. Notably, over 80% of individuals with tertiary education are in HEHC roles, indicating that education provides a significant buffer against displacement by equipping workers with AI-complementary skills. This aligns with findings by Acemoglu and Restrepo (2020), who argue that AI adoption tends to augment high-skill workers while eroding the demand for routine-task occupations.
The IMF report also highlights structural patterns beyond education, such as gender and sectoral concentration. In both countries, female workers are somewhat overrepresented in LE occupations, while male workers are more polarised—clustered in both HELC and HEHC roles. This reflects a broader pattern observed across OECD countries, where women are underrepresented in STEM and digital occupations, and overrepresented in sectors less likely to benefit from AI (OECD, 2023). The result is a gendered exposure to displacement and opportunity. Without deliberate action, AI could deepen existing labour market inequalities. As noted by the ILO (2023), gender gaps in training access and digital literacy pose serious barriers to women’s inclusion in future job markets. Gender-sensitive policy approaches are therefore essential, including affirmative training programs and incentives for employers to diversify recruitment in tech-heavy sectors.
Age also plays a crucial role in shaping the effects of AI on the labour market. Older workers (aged 55 and over), especially those with a college education, are disproportionately found in HEHC occupations—benefiting from their cognitive intensity and lower physical demands. These jobs often offer flexibility, remote work options, and higher wages, aligning with the preferences of ageing workers. However, a vulnerable segment remains: in both the United States and Brazil, 20% to 30% of older workers are still employed in HELC roles, where the risk of displacement is high and the scope for occupational mobility is limited. This low mobility is a central concern. Transition probabilities between occupations decline steeply with age, and workers over 50 face significantly lower chances of switching to less-exposed roles compared to younger cohorts. Without intervention, many older workers in HELC roles may be pushed into early retirement, precarious employment, or unemployment.
The IMF report incorporates the concept of “healthy ageing”, noting that age-friendly jobs—such as those in HEHC categories—can extend working lives. Cognitive health and occupational characteristics interact: jobs that reduce physical strain and offer control over work schedules are especially conducive to older workers’ participation.
To navigate these complex shifts, the IMF outlines several policy directions. First, investment in lifelong learning and re-skilling is critical, particularly for older and less-educated workers. Policy design must recognise the lower transition probabilities of these groups and offer more individualised support. This is consistent with findings by Autor, Mindell, and Reynolds (2022), who emphasise that inclusive technology transitions require systems that make workforce development continuous and adaptive. Second, governments should improve the age-friendliness and gender-responsiveness of jobs by incentivising remote work, promoting flexible hours, encouraging ergonomic job design, and addressing unconscious bias in recruitment. Third, active labour market policies (ALMPs) should target high-risk groups and include job placement services, wage subsidies, and anti-discrimination enforcement.
These micro-level interventions should be embedded within a broader structural reform strategy, as emphasized in the IMF’s macroeconomic outlook. This includes modernising education systems to emphasize AI-relevant skills, reforming social safety nets, and investing in digital infrastructure to prevent regional and digital divides.
AI offers a unique opportunity to boost productivity and economic dynamism—but only if supported by inclusive, anticipatory, and intersectional labour policies. The IMF’s analysis provides a strong empirical basis to guide reforms that align technological progress with social cohesion. Managing the AI transition equitably means acting now—across age, skill, and gender lines.
References
- Acemoglu, D., & Restrepo, P. (2020). Robots and Jobs: Evidence from US Labor Markets. Journal of Political Economy, 128(6), 2188–2244.
- Autor, D., Mindell, D., & Reynolds, E. (2022). The Work of the Future: Building Better Jobs in an Age of Intelligent Machines. MIT Work of the Future.
- International Labour Organization. (2023). The Impact of Artificial Intelligence on the World of Work. Geneva: ILO.
- International Monetary Fund. (2025). World Economic Outlook: A Critical Juncture amid Policy Shifts. Washington, DC: April.
- OECD. (2023). Artificial Intelligence and the Labour Market: What Do We Know So Far? OECD Publishing.
[1] The United States Current Population Survey, and The Brazilian National Household Sample Survey (Pesquisa Nacional por Amostra de Domicílios).
