Send message to Author

Paper short abstract

AI is often seen as a threat to jobs, but its main impact may be widening skill and income inequalities. This paper examines how AI reshapes work, who (given intersectionality of personal characteristics) benefits, and the policy conditions under which AI could reduce rather than deepen inequality.

Paper long abstract

Artificial intelligence (AI) is frequently framed as both a productivity-enhancing tool and a major threat to employment. McKinsey (2025) forecasts 30% of jobs losses in USA by 2030. However, such projections, albeit varied across countries, often assume that the proportion of tasks AI (and robotic) can perform directly translates into job losses. This overlooks how occupations adapt and evolve in response to technological change. Past predictions of occupational displacement—e.g. radiology—illustrate that AI more often reshapes work, augmenting productivity rather than eliminating jobs outright.

This paper argues that the central development challenge posed by AI lies less in aggregate job loss and more in the distributional consequences of AI adoption. AI is more likely to complement workers with higher levels of education, digital skills, and income, while disadvantaging those with fewer opportunities to acquire such skills. Through intersectionality between different personal and household's characteristics, inequalities linked to gender, ethnicity, and socioeconomic status could be worsened. AI risks widening gaps in wages, employment opportunity and job security, with a stronger effects within each country rather than between high-, middle-, and low-income contexts.

We propose a framework grounded in firms and labour market incentives to explain how AI adoption interacts with skill formation and task/job reallocation. The framework highlights conditions under which AI could narrow skill gaps rather than deepen inequality. Using cross-country survey data on AI accessibility and usage, we assess whether these equalising conditions are emerging and identify policy priorities for leveraging AI as a tool for inclusive and equitable development.