In a letter published Monday, **Andrew Bailey**, chair of the **FSB**, underscored the urgent need for preparedness against advanced AI models. These models have already demonstrated capabilities for unauthorized activities targeting third-party systems, as revealed by cybersecurity evaluations at leading AI developers like **OpenAI**, **Anthropic**, and **Meta**, alongside the **UK’s AI Security Institute**. Security agencies are increasingly concerned that AI could significantly reduce the time and cost associated with discovering and exploiting vulnerabilities. This echoes a previous warning from Britain’s **National Cyber Security Centre (NCSC)**, which highlighted increased operational risks for organizations unable to keep pace with an accelerated patching cycle. While companies such as **Microsoft** have shown an observable increase in patch releases, widespread exploitation of newly disclosed flaws by AI has not yet been extensively documented. Bailey's letter, released ahead of a G20 meeting in Asheville, North Carolina, emphasized that the financial system’s heavy reliance on highly concentrated third-party service providers amplifies the danger of system-wide disruption. The **FSB** stressed the critical importance of robust response and recovery capabilities, including the ability to restore critical systems and data from 'bare metal' after a significant cyber incident. Britain's latest annual **National Risk Register** outlines a dire worst-case scenario: a sophisticated cyberattack targeting financial infrastructure that overwrites data on hard drives. In such an event, restoring systems could take years, and a loss of confidence in account balances and transaction records could trigger widespread withdrawals, threatening the stability of the entire financial system. The **FSB** is also exploring how financial services companies can safely deploy frontier AI models for defensive purposes. However, Bailey cautioned that advancements in AI capabilities must be mirrored by stronger resilience and preparedness across the board. A significant concern raised by the **FSB** is the global lack of adequate safeguards governing the development, release, and deployment of advanced AI models. Bailey called for this gap to be a global priority, advocating for coordinated international measures to ensure the safe and responsible release of these powerful technologies. The broader economic implications of AI remain uncertain. While it has spurred a massive surge in technology investment, with AI-related spending estimates ranging from 1.8% to 5% of U.S. gross domestic product, analysts note that evidence of widespread productivity gains beyond this investment boom remains limited.