Deputy Governor Sarah Breeden outlines a new supervisory strategy to address the systemic risks posed by agentic artificial intelligence in financial markets.

The Bank of England is moving to overhaul its supervisory frameworks to account for the rapid proliferation of agentic artificial intelligence within the financial sector. Sarah Breeden, deputy governor at the Bank of England, articulated this shift during a central banking conference in Portugal on July 1, 2026, noting that existing regulatory structures are increasingly ill-equipped to manage autonomous systems.

The central concern centers on the integration of agentic AI into critical infrastructure, specifically regarding automated payments and high-frequency trading environments. Breeden emphasized that traditional oversight models rely on the assumption of constant human intervention, a premise she argues is no longer tenable in an era of autonomous decision-making. The current regulatory architecture lacks the capacity to monitor agents that operate independently of direct human oversight.

Beyond the operational mechanics of AI, the Bank of England is prioritizing the mitigation of heightened cyber threats that accompany these technological advancements. Breeden warned that the frequency and severity of cyber attacks are likely to escalate as AI agents become more deeply embedded in institutional workflows. Policy frameworks must evolve to accommodate these shifts, requiring internal institutional skills that can adapt to rapid technological surprises.

To address these systemic vulnerabilities, the Bank of England is exploring the implementation of advanced scenario analysis and digital twin simulations to model complex interactions between financial institutions. These digital replicas would allow regulators to test how autonomous agents might behave during periods of market stress. By simulating these interactions, the central bank aims to identify potential failure points before they manifest in real-world trading environments.

The proposal also includes the development of enhanced recovery protocols designed to maintain continuity during major system disruptions. Under this framework, the Bank of England is evaluating mechanisms that would allow one financial institution to assume the core processes of another during a technical collapse. This contingency planning reflects a broader effort to ensure that the financial system remains resilient even when individual components fail due to model errors.

Breeden further suggested the adoption of market-wide circuit breakers and automated kill switches to prevent cascading failures triggered by faulty AI models. These tools would serve as a final layer of protection to halt trading activity if an autonomous agent initiates a market meltdown. Such measures are intended to provide a necessary safety buffer in an environment where human reaction times are insufficient to counter algorithmic volatility.

This regulatory pivot aligns with recent international efforts to establish guardrails for artificial intelligence in global finance. The Financial Stability Board recently issued a call for more stringent safeguards regarding the deployment of AI agents, highlighting the cross-border nature of these risks. The alignment between the Bank of England and international bodies underscores the urgency of creating a standardized approach to autonomous financial systems.

The shift toward bespoke regulation represents a departure from traditional, reactive policy-making. By focusing on the unique characteristics of agentic AI, the Bank of England is attempting to balance the benefits of automation with the necessity of systemic stability. The focus remains on building institutional structures that can withstand the unpredictable nature of autonomous agents while maintaining the integrity of the broader financial ecosystem.

Market participants should anticipate a period of increased scrutiny regarding their internal AI governance and risk management practices. The Bank of England will likely prioritize the development of technical standards that define how autonomous agents must interact with core financial systems. Future regulatory guidance will likely focus on the accountability of firms for the actions of their autonomous agents, regardless of the level of human oversight involved.

The ongoing dialogue between central banks and technology developers will determine the feasibility of these proposed safeguards. As the list of AI-driven applications continues to expand, the Bank of England remains committed to updating its oversight tools to prevent technology-induced instability. The coming months will likely see the release of more detailed proposals concerning the implementation of digital twins and automated circuit breakers.

At Global Business & Economics Journal, we contribute to the  future of business and technology with news,  innovative research, expert insights, and thought-provoking analysis.