Artificial intelligence, services globalisation and income inequality

BIS Working Papers  |  No 1135  | 
25 October 2023



Artificial intelligence (AI) is increasingly embedded into modern economies. This prompts fundamental questions about its implications for income distribution. This study seeks to explore how AI-related investment is associated with changes in income and income shares for various segments of the population. As AI continues to expand its capabilities, outperforming human capabilities in an array of tasks, the implications of this technological evolution for income inequality become even more important. The primary focus is on how AI-induced structural shifts in production and the labour market relate to income disparities.


Drawing from a rich data set that covers 86 countries over 2010–19, this paper presents high-level cross-country evidence on how AI-related investment is associated with economic outcomes, including real incomes and income shares across different income groups. We assemble a novel data set combining several public and private sector sources to gain insights into AI investment, real incomes and income shares and labour market outcomes. Our research sheds light on the potential mechanisms through which AI adoption might influence income distribution, including structural transformation, market concentration and skill-biased technical change.


Our results indicate a clear association between greater AI investments and higher income inequality during our sample period. In particular, we find that higher AI investment is associated with higher income and a higher income share for the top decile, while the income share declines for the bottom decile. Investment in AI in the real estate, network technology and robotics sectors has an especially pronounced link with inequality. Furthermore, AI investment is associated with broader economic shifts, including increased total factor productivity (TFP) and modern service exports. There is also a shift from mid-skill jobs to high-skill and managerial positions, accompanied by a decline in the labour share of income. Overall, these findings are consistent with structural transformation and skill-biased technical change, as seen over the last decade, leading to widening income inequality. 


How does economic activity related to artificial intelligence (AI) impact the income of various groups in an economy? This study, using a panel of 86 countries over 2010–19, finds that investment in AI is associated with higher income inequality. In particular, AI investment is tied to higher real incomes and income shares for households in the top decile, while households in the fifth and bottom decile see a decline in their income shares. We also find a positive association with exports of modern services linked to AI. In labour markets, there is a contraction in overall employment, a shift from mid-skill to high-skill managerial roles and a reduced labour share of income.

JEL classification: D31, D63, O32

Keywords : artificial intelligence, automation, services, structural shifts, inequality