Bank of England reviews financial regulations to address risks from autonomous agentic AI systems

1. Regulatory Review of Agentic AI

The Bank of England is evaluating whether current financial regulations are sufficient to govern the use of agentic AI. Deputy Governor Sarah Breeden noted that existing frameworks were not designed for autonomous systems capable of making decisions and executing tasks without direct human instruction. As these systems become more prevalent in areas such as trading, payments, and operational workflows, the Bank of England is considering whether traditional human-oversight requirements remain practical. A 2026 report from the Cambridge Centre for Alternative Finance indicated that 52% of surveyed financial firms are already adopting agentic AI, prompting regulators to examine the need for new guardrails, kill switches, and circuit breakers to prevent market volatility.

2. Cyber Resilience and Stability Risks

The Bank of England has identified cyber resilience as a primary financial stability concern regarding agentic AI. While AI tools can enhance security, officials warn that malicious actors could use similar technology to launch sophisticated, large-scale cyberattacks. The International Monetary Fund (IMF) has echoed these concerns, noting that the interconnected nature of modern financial infrastructure—including shared cloud services and payment networks—could lead to correlated failures if multiple institutions are targeted simultaneously. In response, the Bank of England is exploring stronger recovery requirements, such as improved failover systems and protocols that would allow one institution to assume the essential functions of another during a systemic outage.

3. Global Governance Efforts

International regulatory bodies are increasingly focused on the risks posed by autonomous financial systems. The Financial Stability Board (FSB) recently proposed 12 sound practices for the responsible adoption of AI, emphasizing the need for clear governance and defined responsibilities within financial institutions. While these practices are not binding, they reflect a global push to address the challenges AI poses to human oversight.