Could AI Turn an Ordinary Downturn into an Economic Crisis?

Gita Gopinath, IMF First Deputy Managing Director, recently addressed the AI for Good Global Summit in Geneva, Switzerland, delivering a compelling speech about the transformative potential of AI and its associated risks. While AI is celebrated for its ability to drive productivity and economic growth, Gopinath cautioned that the same technology could exacerbate economic downturns, potentially turning them into severe crises.

Unseen Risks of AI

The focus on AI's risks has traditionally centered on security, privacy, misinformation, and ethical concerns. However, Gopinath highlighted a less discussed risk: AI's potential to amplify economic recessions. She explored how AI could impact labor markets, financial markets, and supply chains during downturns.

AI and Labor Markets

AI threatens to significantly disrupt labor markets, particularly during economic downturns. Historically, firms have invested in automation during good times and laid off workers to cut costs during downturns. AI could accelerate this trend, with 30% of jobs in advanced economies at high risk of AI-driven substitution, compared to 20% in emerging markets and 18% in low-income countries. The result could be unprecedented job losses and long-term unemployment, especially for workers lacking the skills needed in an AI-dominated economy.

AI in Financial Markets

The financial sector's early adoption of AI, particularly in algorithmic trading and robo-advisory services, introduces new risks. AI models, particularly those using machine learning, may struggle during novel economic events, potentially exacerbating market volatility. Gopinath warned that AI-driven trading systems might react overly conservatively in a downturn, leading to a self-reinforcing cycle of asset price collapses.

AI in Supply Chains

AI's role in managing supply chains is another concern. Businesses increasingly rely on AI for inventory and production decisions, which could backfire during a downturn. AI models trained on outdated data might make erroneous decisions, leading to significant supply chain disruptions, shortages, and inflated costs.

Policy Recommendations to Mitigate Risks

Gopinath emphasized the need to manage AI integration carefully rather than curtailing its development. She suggested three key policy actions:

1. Balanced Tax Policies: Ensure tax systems do not inefficiently favor automation over human labor. Current tax policies in several countries favor automation, potentially exacerbating job losses.
2. Workforce Education and Support: Invest in education and reskilling programs to prepare workers for AI-driven changes. This includes enhancing social safety nets with more generous unemployment benefits to help workers adapt to job market changes.
3. Regulation and Early Warning Systems: Strengthen financial regulation and supervision to mitigate AI-related risks. This includes enhancing disclosures, stress testing AI models against unprecedented events, and ensuring sufficient human oversight.

Leveraging AI for Mitigation

Interestingly, AI itself can help mitigate these risks. AI can improve tax compliance, enhance social assistance targeting, and assist in financial supervision by detecting vulnerabilities early. By improving the performance of AI models in novel situations, the risks to financial and supply chain stability can be reduced.

Conclusion

As Gopinath noted, there is an urgent need to "AI-proof" the global economy. This involves closely tracking AI development and adoption, creating and analyzing potential economic scenarios, and making significant investments in understanding AI's impact on the economy. Policymakers must act now to prevent AI from turning the next economic downturn into a crisis while harnessing its potential for good.

AI has the power to change our lives and the global economy. We have the power to shape that change for the better. The future of AI in finance and the broader economy is promising but requires careful navigation to ensure that its adoption does not inadvertently lead to deeper economic crises.