AI could become the next financial shock: IMF tells central banks to prepare before crisis hits
The IMF also encouraged regular stress testing of AI systems, stronger operational resilience standards and contingency planning to ensure financial institutions can continue functioning even if AI systems fail or are compromised.
- Jul 28, 2026,
- Updated Jul 28, 2026 5:30 AM IST
Artificial intelligence is rapidly transforming the global financial system, helping banks process loans faster, improving fraud detection and powering automated trading. But the same technology that promises greater efficiency could also make financial crises spread faster and become harder to contain, according to a new International Monetary Fund (IMF) blog.
In its July 23 blog titled How Central Banks Can Contain Financial Stability Risks as AI Accelerates Change, the IMF says AI is no longer just a productivity tool for financial institutions — it is becoming a systemic force that could reshape the way markets react during periods of stress.
While AI offers significant benefits under normal conditions, the IMF warns that widespread adoption without adequate safeguards could amplify market volatility, increase cyber risks and create new vulnerabilities across the financial system.
AI is making finance smarter
Banks, insurers, asset managers and central banks are increasingly deploying AI across a range of activities, from algorithmic trading and credit assessments to risk management and regulatory supervision.
According to the IMF, these applications can improve efficiency, reduce costs and enhance decision-making. However, they also introduce new forms of systemic risk because many institutions may end up relying on similar AI models, datasets and cloud infrastructure.
If multiple firms use comparable AI systems, they may react to market events in similar ways, triggering synchronized buying or selling that could magnify price swings instead of stabilizing markets. This "herding" effect may cause financial shocks to spread much faster than regulators are accustomed to handling.
Three major risks central banks need to watch
The IMF identifies three broad areas of concern.
- First, correlated market behaviour: AI-driven trading strategies could lead multiple institutions to make similar decisions simultaneously, accelerating market movements during periods of stress.
- Second, concentration risk: Much of the financial sector's AI capability depends on a relatively small number of cloud service providers and AI model developers. Any disruption, cyberattack or operational failure affecting one of these providers could ripple across banks and financial markets worldwide.
- Third, cyber threats: AI is making cyberattacks faster, cheaper and more sophisticated. Attackers can now identify vulnerabilities more quickly, increasing the likelihood of coordinated attacks on payment systems, banks and financial market infrastructure.
What central banks should do
Rather than slowing AI adoption, the IMF argues that regulators should focus on making the financial system more resilient.
The blog recommends that central banks strengthen governance around AI use, improve oversight of institutions deploying advanced models and require greater transparency over how AI systems influence financial decisions.
Regulators should also develop better tools to monitor common AI exposures across the financial system, ensuring they can detect emerging risks before they become systemic.
The IMF further stresses the need for international cooperation, noting that financial markets, cloud infrastructure and AI providers operate across borders. Effective supervision will therefore require regulators to share information and coordinate responses during periods of market stress.
Human oversight remains essential
A recurring message throughout the IMF's analysis is that AI should complement—not replace—human judgment.
The fund cautions against excessive dependence on opaque AI models for critical financial decisions, arguing that human oversight remains essential, particularly during crises when models trained on historical data may behave unpredictably.
The IMF also encouraged regular stress testing of AI systems, stronger operational resilience standards and contingency planning to ensure financial institutions can continue functioning even if AI systems fail or are compromised. The IMF said that AI has the potential to make financial systems more productive and efficient, but only if innovation is matched with robust governance.
Artificial intelligence is rapidly transforming the global financial system, helping banks process loans faster, improving fraud detection and powering automated trading. But the same technology that promises greater efficiency could also make financial crises spread faster and become harder to contain, according to a new International Monetary Fund (IMF) blog.
In its July 23 blog titled How Central Banks Can Contain Financial Stability Risks as AI Accelerates Change, the IMF says AI is no longer just a productivity tool for financial institutions — it is becoming a systemic force that could reshape the way markets react during periods of stress.
While AI offers significant benefits under normal conditions, the IMF warns that widespread adoption without adequate safeguards could amplify market volatility, increase cyber risks and create new vulnerabilities across the financial system.
AI is making finance smarter
Banks, insurers, asset managers and central banks are increasingly deploying AI across a range of activities, from algorithmic trading and credit assessments to risk management and regulatory supervision.
According to the IMF, these applications can improve efficiency, reduce costs and enhance decision-making. However, they also introduce new forms of systemic risk because many institutions may end up relying on similar AI models, datasets and cloud infrastructure.
If multiple firms use comparable AI systems, they may react to market events in similar ways, triggering synchronized buying or selling that could magnify price swings instead of stabilizing markets. This "herding" effect may cause financial shocks to spread much faster than regulators are accustomed to handling.
Three major risks central banks need to watch
The IMF identifies three broad areas of concern.
- First, correlated market behaviour: AI-driven trading strategies could lead multiple institutions to make similar decisions simultaneously, accelerating market movements during periods of stress.
- Second, concentration risk: Much of the financial sector's AI capability depends on a relatively small number of cloud service providers and AI model developers. Any disruption, cyberattack or operational failure affecting one of these providers could ripple across banks and financial markets worldwide.
- Third, cyber threats: AI is making cyberattacks faster, cheaper and more sophisticated. Attackers can now identify vulnerabilities more quickly, increasing the likelihood of coordinated attacks on payment systems, banks and financial market infrastructure.
What central banks should do
Rather than slowing AI adoption, the IMF argues that regulators should focus on making the financial system more resilient.
The blog recommends that central banks strengthen governance around AI use, improve oversight of institutions deploying advanced models and require greater transparency over how AI systems influence financial decisions.
Regulators should also develop better tools to monitor common AI exposures across the financial system, ensuring they can detect emerging risks before they become systemic.
The IMF further stresses the need for international cooperation, noting that financial markets, cloud infrastructure and AI providers operate across borders. Effective supervision will therefore require regulators to share information and coordinate responses during periods of market stress.
Human oversight remains essential
A recurring message throughout the IMF's analysis is that AI should complement—not replace—human judgment.
The fund cautions against excessive dependence on opaque AI models for critical financial decisions, arguing that human oversight remains essential, particularly during crises when models trained on historical data may behave unpredictably.
The IMF also encouraged regular stress testing of AI systems, stronger operational resilience standards and contingency planning to ensure financial institutions can continue functioning even if AI systems fail or are compromised. The IMF said that AI has the potential to make financial systems more productive and efficient, but only if innovation is matched with robust governance.