The International Monetary Fund has called on central banks and financial regulators to strengthen governance frameworks for Artificial Intelligence.

The Fund warned that the rapid adoption of the technology across the financial system could create new systemic risks if left inadequately supervised.

The recommendation was made by the IMF’s Financial Counsellor and Director of the Monetary and Capital Markets Department, Tobias Adrian.

Adrian made the call in a blog post outlining how AI is transforming financial markets, lending, supervision and risk management.

He noted that AI is increasingly being used to price financial risks, allocate credit, execute trades and support supervisory activities.

According to him, this is creating opportunities for greater efficiency while introducing new vulnerabilities that regulators must address.

Adrian identified three immediate priorities for policymakers.

These are strengthening oversight of AI-driven trading, lending and supervisory technology (SupTech); improving transparency around AI adoption, model dependencies and correlated investment strategies; and expanding international cooperation on cyber security and operational resilience.

According to him, AI has compressed the speed of financial transactions, allowing trading, lending decisions and supervisory analysis to occur in real time.

He said that while these innovations have improved market efficiency, they also increase the speed at which financial shocks can spread across institutions and markets.

Adrian said AI-driven trading has enhanced liquidity, lowered transaction costs and improved price discovery under normal market conditions.

He added that AI-powered lending models have strengthened fraud detection and broadened access to credit by incorporating alternative data in assessing borrowers, particularly small businesses and consumers.

However, he warned that these benefits could quickly become sources of instability during periods of market stress.

He explained that as more financial institutions rely on similar AI models, they may respond simultaneously to identical market signals, amplifying volatility.

Citing IMF research, Adrian said some AI-managed investment funds rebalance portfolios significantly faster than traditional funds, increasing the likelihood of synchronized trading that could intensify market swings.

He cautioned that future financial disruptions may arise less from programming errors and more from multiple AI systems reaching similar conclusions and executing comparable strategies at the same time.

To mitigate these risks, Adrian urged regulators to strengthen stress-testing frameworks, improve monitoring of AI-driven investment strategies and collect more comprehensive data on AI adoption, market exposures and model dependencies.