AI at the IMF Spring Meetings Balancing Growth and Governance

The IMF Spring Meetings this year placed artificial intelligence at the center of global economic debate. Across panels, ministerial discussions, and closed-door policy sessions, AI was framed not just as a technological shift but as a macroeconomic force with the potential to reshape productivity, labor markets, and financial systems. The conversations reflected both optimism and caution. While the upside for growth is significant, the risks around governance, inequality, and systemic stability are becoming harder to ignore.

AI as a Macroeconomic Opportunity

A dominant theme throughout the meetings was AI’s potential to boost global growth and improve state capacity. Faster adoption of AI technologies was repeatedly described as a meaningful upside scenario for the global economy. In advanced economies, AI is expected to drive efficiency gains across sectors such as finance, healthcare, and manufacturing. In emerging markets, the opportunity lies equally in leapfrogging traditional development pathways through digital infrastructure and AI-enabled services.

IMF officials emphasized that AI could significantly enhance government capabilities. From improving tax compliance to strengthening public service delivery, AI is increasingly seen as a tool that can help states operate more effectively. This is particularly relevant for countries with limited administrative capacity, where AI could help close long-standing governance gaps. At the same time, discussions made clear that the growth benefits of AI will depend heavily on access.

Labor Markets Under Pressure

Alongside growth, labor market disruption emerged as one of the most pressing concerns. AI’s impact on employment was framed as both an opportunity and a risk. On one hand, automation and augmentation can increase productivity and create new categories of jobs. On the other, they can displace workers in routine and middle-skill roles at a scale that may be difficult to absorb in the short term.

Speakers across sessions highlighted that the pace of change is likely to be faster than previous technological transitions. This raises important policy questions around how economies can adapt. Reskilling and upskilling were consistently identified as critical priorities, alongside reforms to education systems and labor market policies.

There was also recognition that the effects of AI on employment will vary significantly across regions. Advanced economies may experience more immediate disruption due to higher levels of automation, while developing economies could face different challenges related to job creation and informal labor markets. In both cases, the ability of governments to respond quickly and effectively will be a key determinant of outcomes.

Governance Moves to the Centre

There was broad agreement that existing regulatory frameworks are not fully equipped to deal with the scale and complexity of AI, but also a clear reluctance to create entirely new rulebooks. Instead, the emerging approach is to adapt and extend current systems. Discussions focused on strengthening governance around data, model risk, and third-party dependencies, rather than introducing standalone AI regulations. This reflects a growing consensus that many of the risks associated with AI are extensions of familiar challenges, including data privacy, operational risk, and market conduct.

A key priority is the need for clearer accountability within institutions. As AI systems become more complex and autonomous, it is increasingly important to define who is responsible for their outcomes. This includes establishing governance structures that can oversee the development, deployment, and monitoring of AI systems over time.

Transparency and explainability were also recurring themes. Policymakers stressed that AI-driven decisions, particularly those affecting consumers or financial markets, must be understandable and auditable. Without this, trust in both institutions and markets could be undermined.

Financial Services Building Guardrails Around AI

In financial services, the conversations were notably pragmatic. There was little appetite for restricting the use of AI outright. Instead, the focus was on building guardrails that allow innovation to continue while managing risk.

The approach gaining traction is one of risk-based governance. High-impact use cases such as credit decisions, trading strategies, fraud detection, and customer advice are expected to face stricter oversight. Lower-risk applications, particularly those used internally, are likely to be subject to lighter controls. This proportional approach is increasingly seen as the most practical way to balance innovation and regulation.

Another important development is the extension of traditional model risk management frameworks to AI systems. Financial institutions are being encouraged to treat AI models, including generative and adaptive systems, as part of their existing risk infrastructure. This includes maintaining inventories of AI systems, conducting regular reviews, and implementing lifecycle controls.

Data governance has also become a central pillar of AI oversight. Discussions highlighted that poor data quality, weak privacy safeguards, and inadequate consent mechanisms are among the biggest sources of AI-related risk. Strengthening these areas is therefore critical to ensuring that AI systems function reliably and ethically.

The role of third-party providers was another area of focus. As financial institutions increasingly rely on external AI models and cloud infrastructure, regulators are paying closer attention to concentration risk and dependency on a small number of providers. This is likely to remain a key area of regulatory scrutiny going forward.

Singapore’s approach was frequently referenced as a practical model. Its FEAT principles, which focus on fairness, ethics, accountability, and transparency, provide a structured framework for governing AI in financial services without imposing overly rigid rules. More recent guidance in Singapore also emphasizes governance structures, risk management processes, and standards for development and deployment, reinforcing a comprehensive but flexible approach.

Diverging Priorities Between Regulators and Industry

One of the more nuanced dynamics at the meetings was the divergence in priorities between regulators and financial institutions. While both sides recognize the importance of AI, their perspectives are shaped by different objectives.

Regulators are primarily focused on financial stability, consumer protection, and systemic risk. From this perspective, AI is seen as an amplifier of existing risks. Issues such as biased decision-making, data misuse, and correlated model failures at scale are of particular concern. Supervisors are therefore inclined to prioritize control, transparency, and resilience.

Financial institutions, by contrast, are operating under strong commercial pressures. They are looking to use AI to reduce costs, improve efficiency, and enhance customer experience. For many firms, the main challenge is not regulation itself but uncertainty around how existing rules apply to new technologies. As a result, there is a strong demand for clearer and more consistent guidance across jurisdictions.

This divergence is shaping the policy debate. Regulators tend to favor principles-based and technology-neutral approaches, while firms are seeking predictability and proportionality. Despite these differences, there are signs of convergence. Both sides increasingly support risk-based frameworks that apply stricter controls to high-impact use cases while allowing flexibility elsewhere.

Inclusion and the Risk of a Widening Gap

A critical concern running through the meetings was the risk that AI could deepen global inequalities. Several sessions, particularly those focused on Africa and low-income economies, highlighted the danger of uneven access to AI technologies. Participants stressed that countries should not slow down investment in AI despite global economic uncertainty. Instead, the focus should be on building the foundations needed to participate in the AI economy. This includes investing in digital infrastructure, developing local talent, and strengthening governance frameworks.

There was also a strong emphasis on the importance of local data and sovereign AI capacity. Relying entirely on external technologies and datasets could limit the ability of countries to shape AI systems in ways that reflect their own economic and social priorities. Building domestic capabilities is therefore seen as essential to ensuring that AI benefits are broadly shared.

A Converging Global Approach

Taken together, the discussions at the IMF Spring Meetings point to an emerging global approach to AI governance. Rather than creating entirely new regulatory systems, policymakers are focusing on adapting existing frameworks to address AI-related risks. This includes extending model risk management, strengthening data governance, and enhancing oversight of third-party providers.

At the same time, there is growing recognition that governance must be proportionate. Not all AI systems carry the same level of risk, and regulatory approaches need to reflect this. Risk-based frameworks are therefore becoming the default, with more stringent requirements for high-impact applications and greater flexibility for lower-risk uses. Importantly, governance is increasingly being seen as an enabler of innovation rather than a constraint. By providing clarity and reducing uncertainty, well-designed frameworks can help institutions adopt AI more confidently and at scale.

Conclusion

The IMF Spring Meetings made clear that AI will be a defining force in the global economy over the coming decade. Its potential to drive growth, enhance state capacity, and transform industries is significant. However, these benefits are not guaranteed.

The central challenge for policymakers and institutions is to ensure that AI is deployed in a way that is both effective and inclusive. This requires governance frameworks that are robust but flexible, capable of managing risk without stifling innovation.

Ultimately, the message from the meetings was straightforward. AI can be a powerful engine of growth, but only if it is accompanied by the right safeguards. Without them, the technology risks reinforcing the very inequalities it has the potential to reduce.

Sources

1. International Monetary Fund, Spring Meetings 2026 discussions and policy panels on artificial intelligence
2. Statements and remarks by Kristalina Georgieva on AI and global growth
3. IMF analytical work on AI, labor markets, and financial stability
4. Bank for International Settlements guidance on AI, model risk, and financial sector governance
5. Monetary Authority of Singapore, FEAT Principles and AI governance guidance for financial institutions