Urgent focus on cyber controls needed as Bank of England and FCA take action on AI
The Bank of England has confirmed it will test the impact of AI trading agents on market dynamics, including the risk of ‘herding’ behaviour amplifying market volatility, while the FCA has committed to sharing best practice, and expanding its AI Live Testing service. RSM UK says these moves reflect growing concern that advanced AI could materially alter both market behaviour and the scale of cyber risk facing the financial services sector.
Erin Sims, financial services senior analyst at RSM UK said: “Rapid advancement of AI in financial services is driving a much-needed shift in regulatory emphasis away from innovation towards financial stability, resilience and cyber security. With AI capabilities advancing faster than governance frameworks, firms can’t afford to take a ‘wait and see’ approach. AI risks must be viewed through the same lens as operational resilience, systemic risk and critical third-party dependence. Boards and senior management must act now to clarify ownership of AI risk, ensure AI use is inventoried and governed, and test how AI tools behave under stress or failure.
“There also needs to be a renewed focus on cyber and incident response plans, ensuring they realistically reflect the potential for AI-assisted attacks or disruption, particularly in the current uncertain geopolitical landscape. Organisations that focus early on defensive capability, transparency and control will be better placed to withstand scrutiny and respond to escalating cyber threats.
“Firms should expect greater scrutiny from regulators on how AI is incorporated into existing stress testing, operational resilience mapping, and third-party risk frameworks, particularly where the same technologies or providers are relied on throughout the sector. As AI tools become more embedded in core operations and decision-making, regulators are likely to place increased weight on governance, accountability and explainability, testing not just whether controls exist, but whether firms can demonstrate a clear understanding of how AI behaves in stressed or disruptive scenarios.
“Firms that strengthen AI governance, resilience testing and cyber preparedness now will not only reduce risk but are also better positioned to operate confidently in an increasingly AI-enabled financial system.”