Central banks and financial supervisors are not keeping pace with the private sector when it comes to advanced artificial intelligence, according to a research report published on Tuesday.
The study, authored by the Cambridge Centre for Alternative Finance in collaboration with the Bank for International Settlements, the International Monetary Fund and other multilateral institutions, finds that financial firms are embracing AI more than twice as fast as their regulators. Only two in 10 regulators reported using advanced AI, while 24% of authorities collect data on AI adoption within the industry. Meanwhile, 43% of surveyed authorities indicated they have no plans to begin collecting such data within the next two years.
The research draws on responses from a broad sample of market participants: roughly 350 traditional financial institutions and fintechs, more than 140 AI vendors, and about 130 central banks and financial authorities representing 151 countries. The report warns that this gap in empirical information could undermine supervisory efforts, noting that: "This empirical blind spot may undermine the prevailing optimism. Authorities cannot successfully harness or oversee AI if they are navigating its adoption and risks without hard data."
Earlier in April, Anthropic released a model called Mythos. Cybersecurity specialists have flagged Mythos as presenting possible problems for the banking sector and its legacy technology systems. In response, regulators worldwide have engaged with banks to assess how prepared older technology stacks are for interacting with emerging frontier AI models.
The report cites Mythos specifically as an example of the types of systems that might be able to exploit software vulnerabilities at scale. It cautions that such capabilities could limit the effectiveness of existing governance and human oversight mechanisms designed to manage operational and cyber risk.
By documenting differences in adoption rates and the limited data collection among authorities, the study underscores a potential disconnect between the rapid spread of advanced AI across financial firms and the capacity of supervisors to measure and respond to associated risks.