BIS Says Central Bankers Must Upgrade Capabilities to Keep Up with AI Developments

The Bank for International Settlements ("BIS") recommended that central banks urgently upgrade their capabilities to act as both informed observers and sophisticated users of AI.

In a new report, the BIS examined how central banks and supervisory authorities are leveraging artificial intelligence ("AI") to enhance their policy, regulatory, and analytical functions. The BIS found that AI adoption—particularly in machine learning and generative AI—has accelerated rapidly, far outpacing previous waves of technological change. The organization highlighted that central banks, as early adopters of machine learning, are now expanding AI use across core policy areas to improve data quality, forecasting accuracy, and financial oversight. The BIS emphasized that while AI offers significant potential to strengthen monetary and financial stability frameworks, its effective use requires collaboration, investment in human capital, and sound data governance to ensure accountability and public trust.

The BIS highlighted four areas of enhanced AI use for central bankers:

• Information Collection. BIS found that machine learning techniques are improving data quality and efficiency in identifying outliers across large, granular datasets like derivatives reporting.
• Macroeconomic and Financial Analysis. BIS reported that AI is enhancing real-time forecasting of economic indicators such as GDP and improving sentiment analysis derived from unstructured data, including news and social media. 
• Oversight of Payment Systems. BIS cited Project Aurora, an Innovation Hub initiative that applied machine learning and network analysis to pooled, privacy-protected data, detecting up to three times more complex money-laundering schemes and reducing false positives by 80% compared to traditional rule-based systems.
• Supervision and Financial Stability. BIS explained that AI tools are automating risk identification from unstructured data, strengthening stress testing, and creating early warning indicators. 

The BIS also highlighted significant challenges for central banking and supervision, including: (i) that the opaque nature of machine learning models limits explainability while generative AI can produce inaccurate outputs; (ii) resource and talent constraints—reflecting high infrastructure costs and a shortage of skilled professionals; (iii) data, privacy, and governance concerns—arising from the use of unstructured or personal data and low public trust; and (iv) operational and systemic risks such as dependence on a few major technology providers, heightened cyber vulnerabilities, and potential herding behavior that could amplify market stress.