Cambridge Centre report warns regulators lack data to monitor advanced AI

A significant disparity in the adoption of artificial intelligence has emerged between financial institutions and their supervisors, raising concerns over the global ability to monitor systemic risks.

According to research published on Tuesday, 28 April 2026, by the Cambridge Centre for Alternative Finance, financial firms are adopting AI at more than twice the rate of regulators.

The study, conducted alongside the Bank for International Settlements and the IMF, found that only two in ten regulators report "advanced AI adoption," while many others lack the necessary data to oversee emerging harms.

The report highlights the arrival of "Mythos," a frontier AI model released by Anthropic on 7 April, as a primary example of next-generation technology that could exploit software vulnerabilities at scale.

Experts warn that Mythos possesses agentic capabilities that could bypass existing human governance mechanisms, making it increasingly difficult for authorities to maintain firm accountability.

Harish Natarajan, practice manager at the World Bank, remarked during the report's launch that authorities in emerging economies specifically "lack the data and skills needed to embed AI" into their supervisory frameworks.

Furthermore, the study flagged a "critical third-party risk" due to the industry's heavy dependence on a few providers, with 69% of respondents relying on OpenAI.

This concentration could expose the financial system to pricing shocks or supply disruptions. To counter these vulnerabilities, the report suggests that regulators must adopt their own agentic AI systems to match the sophistication of the firms they oversee.

This follows a period of rapid development where AI has transitioned from a supporting tool to a core driver of financial market outcomes.