The BIS as laboratory and observer

Read the PDF edition (3 pages)

The Bank for International Settlements approaches artificial intelligence from three directions. Its Monetary and Economic Department analyses the economic effects of the technology and its implications for monetary policy and financial stability. The BIS Innovation Hub, whose structure Part F, Chapter IV described, builds prototypes that apply the technology to the tasks of central banks and supervisors. The committees hosted in Basel examine how the technology affects the institutions they supervise or oversee.¹

The most comprehensive statement of the BIS view appeared as Chapter III of the Annual Economic Report 2024, under the title "Artificial intelligence and the economy: implications for central banks". The chapter argued that the technology could raise productivity across the economy, that it would change the functioning of the financial system and that central banks should use it themselves while building the expertise and data infrastructure to do so. It also called for central banks to cooperate in sharing knowledge and tools, a role for which the BIS offered itself as a natural platform.² In the following years, the BIS published further papers on the use of large language models in central banking, on the effect of artificial intelligence on inflation and labour markets and on the governance of AI within central banks. It also set up a community of practice through which central banks exchange experience.³

Several projects of the Innovation Hub apply the technology directly. Project Aurora, whose results were published in 2023, tested whether machine learning and privacy-enhancing technologies could detect money laundering networks more effectively when data from several institutions and countries are combined. Using synthetic payment data, the project found that network analysis across institutions identified considerably more suspicious activity than the monitoring of single banks, while producing fewer false alerts.⁴ Project Gaia used large language models to extract climate-related data from the reports of companies, which supervisors need to assess climate risks in the portfolios of banks. Other projects addressed cyber resilience, the analysis of payment data for financial crime and the use of generative models to classify price data for inflation analysis.⁵

The Basel Committee and the Financial Stability Board have approached the subject from the supervisory side. The Basel Committee published a report on the digitalisation of finance in 2024, which examined the risks that artificial intelligence, cloud computing and the entry of large technology firms create for banks and their supervisors.⁶ The Financial Stability Board published its report on the financial stability implications of artificial intelligence in November 2024. It identified third-party dependencies and concentration among service providers, market correlations from the use of similar models and data, cyber risks and model risk as the main vulnerabilities, while noting gaps in the data available to authorities for monitoring them.⁷ These publications give Basel a voice in a debate that is otherwise dominated by the technology companies of the United States and China and by the legislators in Brussels.

Notes

1. On the organisation of the BIS and the Innovation Hub, see Part F, Chapters I and IV. To be verified. ↩
2. BIS, Annual Economic Report 2024, Basel, June 2024, Chapter III, "Artificial intelligence and the economy: implications for central banks". To be verified. ↩
3. BIS Papers and BIS Working Papers on AI in central banking published from 2024 onwards; BIS Innovation Hub, reports on its AI-related projects. Status to be verified. ↩
4. BIS Innovation Hub Nordic Centre, Project Aurora: the power of data, technology and collaboration to combat money laundering across institutions and borders, May 2023. See also Part E, Chapter III. To be verified. ↩
5. BIS Innovation Hub, Project Gaia reports (2023 and 2024); reports on Project Raven, Project Hertha and Project Spectrum. Names, dates and results to be verified. ↩
6. Basel Committee on Banking Supervision, Digitalisation of finance, May 2024. To be verified. ↩
7. Financial Stability Board, The Financial Stability Implications of Artificial Intelligence, 14 November 2024. To be verified. ↩

For general information only. This article does not constitute legal or tax advice, does not replace advice on an individual case and creates no lawyer-client relationship. It reflects the legal position as of August 2025.