Central banks have been seeking help from former central bankers in reviewing their forecasting abilities. The Riksbank asked Mervyn King, former governor of the Bank of England, to examine their forecasting during the 2010s while the Bank of England recently asked Ben Bernanke, former Fed chair, to assess their methods. Bernanke concluded that central banks have performed poorly at forecasting due to global shocks and outdated systems at the Bank of England.
The use of AI in central banking is gaining traction, with some central banks in emerging nations like Indonesia using AI to monitor public reactions to monetary policy. The ECB has launched the Athena project, which employs AI to help banking supervisors scan documents to spot anomalies. Central banks are also monitoring fintech firms that use AI for credit allocation and investment strategies, posing challenges for regulators in terms of understanding the AI models driving these services.
In the context of the EU AI Act and the rise of quantum computing in finance, central banks need to catch up in utilizing AI tools. The assumption is that AI-driven outcomes should aid, not replace, economists and supervisors. However, obstacles remain such as outdated IT systems, data management capabilities, and a lack of AI skills in central banks. Data is identified as a significant hurdle, with central banks potentially benefiting from accessing high-quality datasets that could provide better insights for monetary policy and fraud detection.
The next generation of central bankers may utilize their supervisory powers to obtain high-frequency datasets from payment companies, start-ups, and financial institutions to inform monetary policy decisions and detect fraud. The relationship between central bank digital currencies and AI is rarely discussed, yet the implementation of digital currencies could generate valuable household financial behavior data for policy adjustments. Central bankers will need to adapt their mindsets and communication techniques to effectively utilize AI in their decision-making processes.
In conclusion, central banks are increasingly turning to AI for forecasting and supervisory purposes. While challenges exist in terms of outdated systems and data management, central banks have the opportunity to leverage AI for more accurate and informed monetary policy decisions. The potential benefits of using AI, combined with advancements in data collection and central bank digital currencies, may reshape the way central banks operate in the future. Adapting to these changes will require a shift in mindset and communication strategies for central bankers.
