Generative AI is reshaping the global economy — Are we ready for the consequences?

William Hynes (IGP), Jo-An Occhipinti and Ante Prodan

Generative AI is often celebrated as the next great productivity leap — a breakthrough capable of transforming industries, driving economic growth, and unlocking new forms of value. But beneath the surface, it’s setting in motion a quiet but profound shift in the structure of global labour markets and the broader economy.

In our recent Bulletin of the World Health Organization (pdf) article, we argue that the way generative AI is being deployed — particularly when left to unregulated market forces — risks triggering a new kind of economic disruption. Unlike past waves of automation that primarily displaced routine manual and routine cognitive labour, generative AI is now targeting non-routine cognitive roles — jobs that have expanded significantly as workers moved up the skill ladder in response to previous technological shifts. These roles form the backbone of the professional middle class and brain economy. 

While many individuals and organisations experience AI as a tool that enhances productivity or augments tasks, what’s often missed is the bigger picture: at scale, across the economy, these efficiencies reduce the need for human labour. Unless matched by a dramatic and unlikely surge in per capita consumption, the result is not a wave of new jobs — but a net decline in demand for workers. And unlike past disruptions, there’s no clear pathway for “upskilling” when generative AI is evolving rapidly and already outperforming humans in a wide range of complex, non-routine cognitive tasks — faster, cheaper, and without fatigue.

This dynamic threatens to set off a self-reinforcing cycle: widespread job displacement, underemployment, falling incomes, suppressed demand, and recessionary pressures — all unfolding in a context where conventional economic tools like fiscal stimulus or monetary easing may no longer work as intended.

The risk is that we reach a tipping point — where AI’s ability to replace human labour vastly outpaces the economy’s ability to create new, meaningful roles.

A fork in the road: reinvest or extract value?

One of the central unknowns in this transition is how organisations will use the productivity gains unlocked by generative AI.

There are two ways productivity gains from generative AI could play out. One sees workforces downsized and value extracted — using AI to cut labour costs, increase margins, and prioritise shareholder returns. The other retains workers and reinvests — translating gains into higher wages, shorter working weeks, and workforce development. Which path is taken will have profound implications for the distribution of economic opportunity in the years ahead.

The scale of potential displacement — and why it matters

While estimates vary widely, recent projections based on the early capabilities of generative AI offer a glimpse into the potential scale of disruption. The Brookings Institution has estimated that around 60% of job tasks in the U.S. are at medium to high risk of being automated over the next decade. A 2023 Goldman Sachs analysis suggests that generative AI could expose up to 300 million full-time jobs globally to automation, particularly in high- and middle-income countries. And according to the World Economic Forum, employer expectations point to a net loss of 14 million jobs globally by 2027 — with 83 million jobs anticipated to be displaced and only 69 million created.

Even where labour is still needed, demand may fall dramatically, as far fewer workers are required to deliver the same or greater outputs. And because many of the roles at risk are middle- and high-income jobs — those that sustain household spending, tax revenues, and broader economic dynamism — their erosion poses a serious threat to the stability and inclusiveness of modern economies.

The U.S. acceleration effect

Under the Trump administration, the United States has reversed course on AI governance, dismantling many of the safeguards introduced in recent years. At the same time, venture capital investment in AI companies has surged, and major tech firms are integrating generative AI into core products and services — from productivity software to customer service — with little incentive to retain human labour in the process.

This deregulatory push, combined with rising economic nationalism and ongoing political instability, is already creating ripple effects well beyond U.S. borders. Without coordinated international action, there is a growing risk that the U.S. model will drive a race to the bottom — weakening global labour standards, encouraging exploitative forms of automation, and fuelling broader economic and geopolitical volatility.

What needs to be done — now

To avoid a future of deepening inequality, economic fragility, and social unrest, governments must do more than react — they must reshape the rules of the game. That means treating generative AI not just as a tool of efficiency, but as a force that demands a new social contract: one where productivity gains are shared, where human contributions are valued beyond those made to the formal economy, and where technological progress is governed in service of collective quality of life. From rethinking taxation and ensuring public stakes in AI infrastructure, to embedding foresight into regulation and investing in social infrastructure, the path forward is not about resisting innovation — but making sure it works for everyone. 

The disruption is here. The window to shape it is closing. 

William Hynes is Honorary Professor of Practice at the UCL Institute for Global Prosperity (IGP). He is a Senior Climate Change Economist and co-Director of the Coalition for Capacity on Climate Action (C3A) Program at the World Bank.

Associate Professor Jo-An Occhipinti is an NHMRC Principal Research Fellow and Co-Director of the Mental Wealth Initiative at the University of Sydney’s Brain and Mind Centre. She is a systems scientist focused on aligning public policy with social, economic, and environmental resilience.

Dr Ante Prodan is a computer scientist and complex systems researcher with the School of Computer, Data and Mathematical Sciences at Western Sydney University. He serves as the Mental Wealth Initiative’s Chief Scientific Advisor on AI, where his research explores the societal impacts and opportunities of generative AI