Frontier AI and cyber resilience: what financial services firms need to know now

Firms using frontier AI models for cyber resilience say that the value they get from them is determined less by the models themselves and more by the technical and operational environment they are deployed in.  A harness is the collection of tools, workflows, controls, data and operating environments that sit around a frontier AI model, and in cyber defence it helps translate model capability into outputs that can be reviewed, validated and acted upon in practice.

Access to a powerful AI model does not automatically create an effective cyber defence capability, a well-designed harness can help turn model capability into cyber defence outputs that are usable, explainable and capable of being acted upon.  This means that the benefits organisations obtain from frontier AI may depend as much on harness design as on the underlying model itself.

Firms can choose from a growing range of ways to build and operate AI harnesses, each with different strengths and limitations, including internally developed tools, vendor-provided platforms, managed third-party offerings and open-source components, and no single approach appears to provide a complete solution at this stage.  Wrapper or orchestration layers can provide a way to coordinate multiple harnesses, reduce early dependency on a single model or vendor, and a modular architecture may help compare capabilities, support substitution between tools and preserve flexibility as frontier models and harness technologies continue to evolve.

One of the practical challenges in applying frontier AI to cyber defence is deciding how much organisational information the model and harness should be allowed to access, model effectiveness is highly dependent on the quality and relevance of organisational context, including source code, asset and environment information, security architecture and control data.  However, deciding what information can be safely provided to frontier AI models and harnesses remains a central risk management question, with relevant considerations including confidentiality, data protection, supplier assurance, information handling restrictions and potential misuse if a powerful model or harness were compromised.

Governance decisions can be reflected through practical controls within the harness and its operating environment, including use-case restrictions, controlled user access, network isolation, environment segregation, monitoring, approval workflows, rules of engagement and controls that limit when and how users or systems can obtain access.  Model-level or platform-level guardrails may therefore be only one part of the control design, with additional controls applied through the harness, user permissions and operating environment.

Frontier AI is changing the speed, scale and manner in which vulnerabilities can be discovered and may need to be addressed, with firms now able to identify weaknesses in software, systems and infrastructure rapidly, fundamentally changing how they view cyber resilience.  Even where a substantial proportion of model outputs are ultimately discounted through expert review, the remaining volume of genuine vulnerabilities can still put considerable pressure on remediation teams, engineering resources and change management processes, and firms may need to understand where bottlenecks will arise and whether existing processes can accelerate without creating operational instability.

Frontier AI models can combine multiple lower-rated security flaws (vulnerability chaining) and create alternative routes to compromise, and the relationships between vulnerabilities, systems and dependencies may not have been visible through traditional approaches to testing and scanning.  This is encouraging firms to have a broader view of cyber risk that looks past severity ratings alone, emphasising a risk-based approach using factors such as exploitability, business service impact, prerequisites to exploit, risk-reduction controls and dependency on the vulnerable system.

Frontier AI is revealing whether firms have the right governance, risk ownership, engineering capacity and remediation processes needed to act on AI-generated vulnerability discovery, organisational readiness is the primary challenge.  As firms use frontier AI to identify vulnerabilities, the importance of asset mapping, dependency management, risk ownership, remediation capacity and triage processes will rapidly become apparent, several firms have characterised frontier AI as a stress test of their existing cyber-resilience capabilities.

Firms highlighted the importance of supplier preparedness, cloud dependencies, software supply chain visibility and shared infrastructure, and some firms are already engaging suppliers on using AI-enabled vulnerability discovery, how they are validating findings, letting customers know, and whether they are able to remediate quickly.  Across the wider supplier ecosystem, some third parties may have fewer engineering resources than larger firms, creating potential dependencies as AI-enabled vulnerability discovery and remediation activity increases.

Harness effectiveness is not only about identifying potential issues, scaling is likely to depend on the capacity to triage, validate, prioritise and remediate outputs without overwhelming engineering, cyber or remediation teams, and on integration with existing engineering and security tooling.  Scaling the use of frontier AI may therefore depend as much on people, processes and remediation capacity as on the AI capability itself.

Although frontier AI is increasing the degree of automation in cyber resilience activities, firms report that human oversight remains critical. models can accelerate vulnerability discovery, code analysis and inform prioritisation, but firms continue to rely on specialist expertise to validate findings, assess relevance, determine priorities and make risk-based decisions.  Governance forums, risk committees and senior leaders may need clearer visibility of how frontier AI affects vulnerability registers, remediation, supplier dependencies, risk and operational resilience, and firms may also need to distinguish between observed risks, firm-specific evidence and more speculative scenarios.