The International Monetary Fund has issued a stark warning about artificial intelligence adoption in financial markets, calling on central banks globally to strengthen oversight of AI systems. According to the IMF's blog published on Thursday, while AI-driven execution can help improve liquidity, lower transaction costs, and accelerate price discovery under normal market conditions, during periods of stress, those same features can amplify risks. The multilateral institution warned that future flash crashes may arise less from coding errors and more from many AI systems reacting in parallel to the same information, creating what it calls 'many-AI' risks. Latest developments show that US banks expect AI to reduce expenses by up to 5%, nearly double the revenue growth projected over the next three years, with larger institutions forecasting even stronger gains.
The IMF's recommendations carry particular significance for India, where the government and regulators are increasingly promoting AI adoption across the financial sector. As reported by the IMF, India is deploying AI-powered credit assessment to help lenders move beyond conventional credit scoring by analyzing digital payment transactions, Goods and Services Tax filings, bank statements and utility payments. According to government estimates, such models could unlock an additional $130-170 billion in economic value by narrowing the credit gap for smaller businesses, including micro, small and medium enterprises (MSMEs), first-time borrowers and informal workers. In the US, banks are currently focusing AI adoption on operational efficiency improvements, with most use cases targeting automated processes such as Bank Secrecy Act compliance and call-centre operations.
Several regulatory initiatives are underway in India to manage AI risks. The Reserve Bank of India's Regulatory Sandbox allows banks and fintech firms to test AI-enabled financial products under regulatory supervision, while the Reserve Bank Innovation Hub has developed MuleHunter.AI to detect mule bank accounts used in cybercrime through AI-driven transaction analysis. However, the IMF cautioned that while AI can help alleviate skill shortages, it requires specialized expertise that is scarce in emerging markets such as India. The institution warned that model risk and over-reliance on automated outputs can create blind spots, especially when systems perform poorly under stress. In the US, banks are prioritizing cost reduction over revenue growth, with most respondents identifying operational efficiency as the main driver for AI investment.
The IMF has identified several critical vulnerabilities in the AI-financial system. According to the institution, the growing dependence on a handful of cloud, data and AI model providers could create new systemic vulnerabilities, with disruptions at a critical service provider potentially affecting multiple financial institutions simultaneously. The IMF also warned that generative AI is increasing the sophistication of cyberattacks, making cyber resilience a broader financial stability concern rather than merely an operational risk. In the US, banks are implementing AI for cost reduction through automation and process streamlining, with the financial impact depending on implementation costs, data quality, model governance, cybersecurity and regulatory compliance. The survey reflects that banks may use AI to limit future staff expansion and redeploy employees towards revenue-generating functions.
The IMF emphasized that AI should augment, rather than replace, supervisory judgement, recommending that financial supervisors would need stronger technical capabilities, governance standards and human oversight to manage increasingly complex AI systems. According to the IMF, if policymakers act early and collectively, AI can reinforce global financial resilience. The institution concluded that if they do not, future instability may be faster, more correlated, and harder to manage than past episodes, highlighting the urgent need for coordinated international action on AI financial oversight. The practical approach to AI adoption in banking suggests that institutions initially focus on measurable efficiency gains before expecting significant revenue growth, with the financial impact depending on implementation quality and regulatory compliance.
