Experts warn that the widespread adoption of AI could turn a regular recession into a deep and prolonged economic catastrophe, with devastating consequences for labour markets, financial systems, and global supply chains
As the world prepares for the inevitable next economic downturn, a new and alarming risk has emerged – the potential for Generative Artificial Intelligence (GenAI) to exacerbate and prolong the crisis. Experts warn that the widespread adoption of AI could turn a regular recession into a deep and prolonged economic catastrophe, with devastating consequences for labour markets, financial systems, and global supply chains.
The extraordinary benefits that Generative AI could bring, from helping us live healthier lives to accelerating scientific breakthroughs. Experts believe the new technology could also provide a major economic boost by enhancing productivity.
However, that AI’s promise comes with considerable uncertainty and risks. While previous warnings have focused on issues like security, privacy, and ethics, a different kind of AI-related risk has received much less attention – the potential for AI to exacerbate economic crises.
Waiting for the next disaster
The next downturn is inevitable, warns IMF First Deputy Managing Director, Gita Gopinath in a speech at the AI for Good Global Summit, Geneva, Switzerland, and the widespread use of AI, she says could make it far worse. Here’s how:
- Labour Market Disruption: Past waves of automation have shown that the true impact of job-replacing technologies only becomes apparent during economic downturns when firms let go of workers to cut costs. AI threatens to accelerate this trend, with an estimated 30% of jobs in advanced economies at risk of being replaced. This could lead to widespread retrenchment and extended unemployment as seen never before, because the workers who will lose their jobs will not have the necessary skills to survive in an AI-driven,new-age economy.
- Financial Market Instability: The financial services industry has increasingly relied on complex, self-learning AI models for trading and investment decisions. While this can improve efficiency in normal times, these AI systems could struggle to respond appropriately in a future downturn characterised by unfamiliar patterns. The “black box” nature of these models would make managing such an event particularly challenging, potentially leading to a self-confirming spiral of fire-sales and collapsing asset prices.
- Supply Chain Disruptions: Generative AI’s user-friendly features could encourage widespread adoption by companies, who would then come to rely on AI predictions for their production decisions. In a downturn, however, these AI algorithms trained on stale information could trigger a series of forecasting errors, causing crippling delays and shortages of critical supplies across the global economy – arepeat of the costly disruptions seen during the COVID-19 crisis.
Collectively, these AI-amplified disruptions could potentially snowball regular downturns into severe and protracted economic crisis, which could pose massive challenges for policymakers. And if this were to happen in the next few years, it would strike countries already dealing with high debt and low growth, seriously constraining their ability to support workers and firms.
Act now to “AI-proof” the economy
So, what can be done to prevent such a catastrophic scenario? Ms Gopinath outlines three key policy actions that should be taken now:
- Reconsider tax systems that favour automation over people: New research shows that the tax systems of several countries tend to make investments in labor-substituting automation more attractive than investments in labor-complementing technologies. Given the economic risks outlined, it’s questionable whether the social returns from such automation are high enough to justify these tax incentives.
- Protect workers by investing in education, training, and social safety nets: Heavier investments in education and digital competencies are essential, especially in emerging markets and developing economies, where young people are ill-prepared for the rapid technological changes. Additionally, strengthening unemployment insurance and exploring new forms of wage insurance can help workers adapt to job market disruptions.
- Enhance financial regulation and supply chain resilience: Financial regulators will need to upskill to better understand AI-related risks and consider measures such as enhanced disclosures by financial institutions, stress-testing of AI models, and establishing sufficient human oversight to prevent cascading breakdowns. Companies across sectors should also stress-test their AI-reliant operations against potential disruptions.
She acknowledges that this agenda may sound daunting, but it also highlights how AI can be harnessed to help implement many of these policy measures. For example, AI can assist in upskilling workers, improving tax compliance and social assistance targeting, and enhancing financial supervision and risk assessment.
The key message is that policymakers need to act now to “AI-proof” the economy, just as they have focused on mitigating concerns around security, privacy, ethics, and disinformation. This will require significant investments in research and analysis to understand how AI might impact various economic variables, so that they are not “flying blind” in the next recession.
Ms Gopinathemphasises that this is not a prediction, but a significant risk that demands urgent attention. Harnessing the immense potential of Generative AI while mitigating its economic threats is a crucial challenge for policymakers worldwide. With the right actions today, we can shape the transformative power of AI for the better and protect our societies from the looming risk of an AI-amplified economic crisis.
Acknowledgement: www.imf.org