The International Monetary Fund is warning that artificial intelligence is becoming a financial-stability issue as banks, investors and supervisors increasingly rely on the same AI models, data and technology providers. In a new analysis, IMF Financial Counsellor Tobias Adrian argues that AI can make the financial system more efficient and inclusive, but also risks amplifying downturns and accelerating market disruptions when many firms react to the same signals at machine speed.
The IMF's concern centers on how AI differs from earlier financial technologies. AI compresses both time and distance: decisions that once took days now happen in seconds, and vast amounts of information can be processed simultaneously across institutions and markets. That acceleration brings real benefits - faster credit decisions, more efficient markets, earlier detection of risks - but it also creates new channels through which disruptions can spread.
"Individual firms may appear well managed, yet the system may become vulnerable because many firms are making similar decisions at the same time," Adrian writes. "The lesson from previous episodes of financial instability is that system-wide risks often emerge not from the failure of a single institution but from collective behavior."
Lending, trading and the risk of synchronized cycles
In lending, AI systems can evaluate borrowers more efficiently, analyze financial statements and incorporate alternative data sources into risk assessments. That could expand credit access for households and small businesses that lack extensive credit histories. But the IMF warns that when many lenders rely on similar models and data, credit standards could tighten simultaneously during downturns, reducing financing precisely when it is needed most.
Explainability is a growing concern. Some advanced AI models operate as "black boxes," making it difficult for managers, boards or supervisors to understand why a particular lending decision was made. This becomes more acute as lending becomes "agentic," the IMF says, with AI systems initiating and executing parts of the lending process themselves rather than merely supporting human credit officers.
AI is also reshaping trading and investment. Investment firms increasingly use machine learning and generative AI to analyze earnings calls, regulatory filings, news reports and market sentiment. Under normal conditions, AI-powered trading improves liquidity and reduces transaction costs. But during stress, many AI-driven systems responding to the same signals could amplify price declines, accelerate liquidity evaporation and increase volatility.
The IMF points to research suggesting that some AI-based investment strategies rebalance positions more rapidly than traditional approaches. In less liquid markets, circuit breakers that stabilize major equity markets may be less effective protections.
Concentration risk and cyber threats at machine speed
Some of the most important risks may come from the infrastructure supporting AI rather than the algorithms themselves. Many financial institutions depend on a small number of cloud providers, data vendors and model developers. A disruption at a critical provider - whether from operational failure, cyberattack or geopolitical tension - could affect many institutions simultaneously.
The IMF notes that several authorities have begun expanding operational-resilience frameworks to include critical third-party service providers. Understanding how institutions are connected through common providers may be as important as understanding their direct financial linkages.
Cybersecurity is the most rapidly evolving risk. Generative AI is changing both sides of the cyber-defense equation: defenders can identify vulnerabilities and respond to incidents faster, but cybercriminals can automate attacks and develop convincing phishing campaigns more quickly. "The result is a race between attackers and defenders that is increasingly measured in minutes rather than months," Adrian writes.
Regulators are also adopting AI themselves. Supervisory technology, or SupTech, can help authorities monitor markets, analyze large datasets and identify emerging vulnerabilities. But the IMF cautions that supervisors must avoid becoming overly dependent on automated outputs. "AI should augment supervisory judgment, not replace it," the report states.
Four principles for keeping AI safe
The IMF's path forward rests on four principles: governance, with clear accountability for AI-driven decisions; transparency, so policymakers can see where AI is used and where systemic dependencies are forming; resilience, preparing for operational disruptions and model failures; and cooperation, since AI-related risks cross national borders.
"The impact of AI on financial stability is not predetermined," Adrian writes. "AI could make the financial system more efficient, inclusive and resilient by improving capital allocation, supervision and risk management. But these outcomes will not emerge automatically."
The IMF's analysis is a useful reference point for AI for Finance professionals, particularly those in roles where model risk and governance fall within their remit. For finance leaders, the report's emphasis on explainability and model limitations has direct operational implications.
Why this matters for finance professionals
For those working in lending, trading or risk management, the IMF's core message is that AI governance is no longer a technology issue - it is a financial-stability issue. The practical takeaway: institutions should test AI models rigorously under adverse conditions, maintain human oversight of automated systems and map their dependencies on shared technology providers before a disruption exposes them. Finance professionals who understand these risks will be better positioned to build AI systems that survive their first real stress test. Those who want to build that expertise can find structured training through the AI Learning Path for CFOs.
Your membership also unlocks:
