Five System-Level Investments Needed to Share AI's Gains Equitably

As artificial intelligence advances at unprecedented speed, a critical question confronts policymakers and business leaders: Will AI create broadly shared prosperity, or will it concentrate wealth in fewer hands?

New research from MIT Sloan School of Management suggests the answer depends on deliberate choices made now. In a working paper, postdoctoral associate Isabella Loaiza and professor Roberto Rigobon examine how societies can ensure AI becomes a general-purpose technology that enhances welfare for many rather than generating returns for a select few.

The researchers point to historical precedents. Automobiles transformed society only after massive investments in roads, traffic laws, safety standards, licensing systems, and insurance frameworks. The internet, by contrast, has seen insufficient investment in governance, public data infrastructure, and equitable access policies—and its gains have been distributed unevenly as a result.

"We're at a point where this wave of AI is nascent, and we can still decide how to deploy it in society so that it brings welfare to many," Loaiza said. "I would not like to see AI increase concentration of resources and higher inequality."

Why it matters

Unlike previous technological shifts that unfolded over decades, AI is advancing rapidly while societies are still determining how to govern it. The window for shaping complementary systems that distribute AI's benefits equitably is narrow. Without coordinated action across five key areas, AI risks exacerbating existing inequalities rather than closing them.

Five categories of system-level complements

Loaiza and Rigobon identify five interconnected categories of investment required to translate AI's technological progress into genuine improvements in worker welfare:

Capital infrastructure. AI requires accessible, usable infrastructure funded through a mix of public and private investment. The division of responsibility for AI infrastructure—and the extent of each sector's commitment—remains undetermined.

Human development. Workers need support transitioning from linear career paths to cyclical models that emphasize continuous reskilling and renewed purpose. This requires robust programs for skill development and safety nets like unemployment insurance to help workers navigate transitions.

Social and economic institutions. New labor standards and tax reforms are necessary to reduce companies' financial incentives to replace human workers with machines that require no wages or benefits. Such governance mechanisms can help preserve worker autonomy.

Metrics and measurement. Companies and governments must track how labor and technology evolve together, how AI's impact varies geographically, and where AI introduces financial or infrastructure risks. Without proper measurement, policy and investment decisions lack grounding.

Systems thinking. AI must be understood as a component within larger social, economic, and technical systems rather than as an isolated tool. Local actions can have global consequences, requiring coordinated approaches to governance.

The highest-impact investment

While all five complements are essential, Loaiza identified workforce development as likely to deliver the biggest impact relative to resources invested. "We need to figure out a way to train people in a way that complements AI," she said. "When we have people who can use the technology, then workers will have jobs to do."

In previous research, Loaiza and Rigobon defined skills that complement AI's limitations, emphasizing human capabilities requiring empathy, presence, opinion, creativity, and hope.

The researchers acknowledge that building system-level complements is challenging, requiring long-term vision and sustained budgets that conflict with quarterly earnings cycles and short election timelines. Yet the alternative—allowing AI to develop without these supporting structures—risks repeating the internet's pattern of concentrated gains.

"Artificial intelligence holds immense potential to reshape our societies, economies, and daily lives, but its trajectory is not predetermined," the authors write in their paper, "From Wealth to Welfare: The Social and Economic Institutions to Complement AI."

The details were first reported by MIT Sloan School of Management.