Could AI Widen the Gap Between Wealthy and Low-Income Populations?

A Real Concern, Not a Certainty

Many economists and policy researchers have raised genuine concern that AI could widen the economic gap between wealthy and low-income populations if current trends continue without meaningful policy intervention. This concern is grounded in observations about how the economic benefits of AI adoption appear to be distributed so far, rather than a definitive prediction about an inevitable outcome. Researchers studying this question generally emphasize that AI’s ultimate effect on inequality is not predetermined by the technology itself, but will be significantly shaped by policy choices made by governments, businesses, and institutions as AI adoption continues.

How the Concern Is Grounded

The basis for this concern includes several observed and analyzed patterns. Economic analyses have suggested that gains from AI adoption so far appear concentrated among owners of capital, highly skilled workers who can effectively leverage AI tools, and the technology companies developing and deploying AI systems. Meanwhile, workers in roles involving more routine or predictable tasks have been identified in various analyses as facing relatively higher exposure to AI-driven automation and job displacement, a pattern that, if it continues without effective policy responses like retraining programs, could disproportionately affect lower-income workers who are more likely to hold such roles.

Unequal access to AI tools themselves is another factor researchers point to: effectively using advanced AI tools often requires reliable internet access, appropriate devices, and specific skills or training, which are not equally distributed across income levels, potentially limiting who can capture AI’s productivity benefits even when the tools are nominally available to everyone.

Why Outcomes Are Not Considered Predetermined

Despite these concerns, researchers studying this area generally stress that history offers examples of previous major technological transitions that had varied effects on inequality depending significantly on the policy environment surrounding them, rather than being determined purely by the technology’s inherent characteristics. Proposed policy responses that researchers and policymakers have discussed include investment in worker retraining and education programs, updated labor market and social safety net policies, and various proposals for ensuring the economic gains from AI are shared more broadly across society, such as through taxation or other redistribution mechanisms. There is no single agreed-upon solution, and views differ on which specific policies would be most effective, but there is broader agreement that policy choices, not just the technology itself, will play a decisive role in shaping the outcome.

Bottom Line

Many researchers believe AI could widen the gap between wealthy and low-income populations if current trends in how AI’s economic benefits are distributed continue unaddressed, since gains appear concentrated among capital owners and highly skilled workers while lower-income workers face greater displacement risk and unequal access to AI tools. However, this outcome is not considered predetermined, and researchers emphasize that policy choices around education, labor protections, and how AI’s gains are shared will significantly shape the ultimate result.

Go deeper

• Are Schools in Different Countries Adopting Classroom AI at the Same Pace?