IMF Warns Central Banks to Tighten AI Oversight Over Financial Stability Risks

The International Monetary Fund (IMF) has urged central banks and financial regulators around the world to strengthen oversight of artificial intelligence (AI), warning that the technology's rapid adoption across the financial system could create new threats to global financial stability.

The Fund said AI is already transforming the way financial institutions assess risk, allocate credit, monitor markets and respond to emerging economic and financial shocks.

However, while the technology is delivering significant gains in efficiency, reducing costs and improving decision-making, the IMF cautioned that weak governance and inadequate safeguards could allow AI-driven activities to amplify market volatility, cyber threats and systemic risks.

The warning was contained in a new IMF Blog authored by Tobias Adrian, the Fund's Financial Counsellor and Director of its Monetary and Capital Markets Department.

Adrian said AI had become increasingly embedded in the decision-making architecture of modern finance, making it essential for central banks, supervisors and financial regulators to ensure that its deployment strengthens rather than undermines financial stability.

The IMF identified three immediate priorities for policymakers: strengthening oversight of AI-driven trading, lending and supervisory technology; improving visibility into how AI is being used and where concentration risks may exist; and deepening international cooperation on cybersecurity, operational resilience and financial stability.

According to the Fund, AI is compressing time and distance across financial markets, with trading decisions, lending assessments and regulatory analysis increasingly taking place in real time.

While this speed can improve efficiency under normal market conditions, the IMF warned that it could also accelerate the transmission of financial shocks when markets come under stress.

AI-powered trading systems, the Fund noted, are already widely used by investment banks, hedge funds and asset managers to analyse earnings reports, regulatory filings, economic indicators and other market information within seconds.

The technology has helped improve liquidity, lower transaction costs and accelerate price discovery. But the IMF cautioned that the growing use of similar machine-learning models across financial institutions could create a new form of systemic vulnerability.

If numerous firms rely on similar models and respond to the same market signals at the same time, the resulting wave of synchronised trading could intensify market movements and deepen volatility.

The Fund said some AI-driven investment funds are capable of rebalancing their portfolios considerably faster than conventional funds, raising the possibility that large numbers of automated systems could respond simultaneously to changes in market conditions.

The IMF warned that future flash crashes may therefore be driven not only by programming errors or isolated technical failures, but also by multiple AI systems independently reaching similar conclusions and executing similar trades at almost the same time.

The issue is further complicated by the lack of transparency surrounding some advanced AI models.

The Fund noted that even sophisticated financial institutions can struggle to explain precisely how certain AI systems arrive at decisions, particularly when markets are under severe pressure.

This lack of explainability, it said, could make it more difficult for regulators and financial institutions to identify the source of a disruption or intervene quickly when an AI-driven strategy begins to generate unexpected outcomes.

To address these risks, the IMF urged regulators to strengthen monitoring of AI-based trading strategies, improve data collection on AI adoption and model dependencies, and incorporate AI-related scenarios into financial stress tests.

The Fund also warned that AI is increasingly becoming part of the operational infrastructure that supports banks, payment systems, stock exchanges and clearing houses.

As financial institutions rely more heavily on AI and cloud-based technologies for risk management, fraud detection, system monitoring and other critical functions, the IMF said concentration risks could become more pronounced.

A major disruption affecting a key technology provider, whether caused by a technical failure, cyberattack or geopolitical event, could potentially affect multiple financial institutions simultaneously if they depend on the same cloud infrastructure, data provider or AI model.

The IMF pointed to steps already taken by authorities such as the European Central Bank and the Bank of England to broaden operational resilience frameworks to cover AI and cloud service providers.

It urged other jurisdictions to develop similar safeguards to ensure that critical financial services can continue operating even when technology providers experience major disruptions.

Artificial intelligence is also transforming the way financial regulators themselves conduct supervision.

The IMF said Supervisory Technology, or SupTech, is allowing regulators to process large volumes of financial and market data, identify unusual patterns and detect emerging risks more quickly.

Several central banks, including those in France, Germany, Portugal and Japan, already use machine-learning tools for market surveillance, while the United States Federal Reserve, European Central Bank and Bank of Canada use natural-language processing to analyse supervisory reports and consumer complaints.

Despite the growing role of AI in financial supervision, Adrian stressed that technology should complement rather than replace human judgment.

The IMF warned that excessive dependence on automated systems without sufficient human expertise and strong governance could create new blind spots, especially during periods of financial crisis when unusual circumstances may fall outside the parameters on which AI models were trained.

Cybersecurity is another area where the Fund sees growing risks.

The IMF said generative AI is making phishing attacks increasingly convincing, allowing fraudsters to develop more sophisticated schemes and adapt them in real time.

The technology is also shortening the period between the discovery of software vulnerabilities and their potential exploitation, increasing the pressure on financial institutions and regulators to detect and respond to cyber threats quickly.

The Fund therefore called on central banks and financial institutions to strengthen cyber resilience, conduct stress exercises specifically designed around AI-related threats and invest in defensive AI technologies capable of identifying and countering increasingly sophisticated attacks.

For the IMF, the growing use of artificial intelligence has moved beyond being simply a technological or innovation issue. It has become a financial stability concern that cuts across financial markets, individual institutions, critical infrastructure and the regulatory system.

The Fund said policymakers would need to act early to ensure that the benefits of AI do not come at the expense of financial stability.

Adrian also stressed the importance of international cooperation, given the global nature of financial markets and the fact that technology providers, financial institutions and data infrastructure often operate across national borders.

He warned that without coordinated action and effective governance, future financial disruptions could spread more quickly and become more closely linked than previous crises.

“If policymakers act early and collectively, AI can reinforce global financial resilience. If they do not, future instability may be faster, more correlated, and harder to manage than past episodes,” Adrian said.

The IMF's warning comes as financial institutions and regulators worldwide accelerate their adoption of artificial intelligence, highlighting the growing challenge of balancing technological innovation with the need to protect the stability and resilience of the global financial system.