UK AI Cyber Security Code of Practice 2.2: Design for resilience against adversarial attack, unexpected input and failure

What does 2.2 require?

Developers and system operators design and build AI systems so that they hold up under adversarial AI attacks, inputs the system was never meant to see, and failure of the AI component itself.

Maintained by Gerard Blokdyk

Other controls in Secure design (Principles 1 to 4) – UK AI Cyber Security Code of Practice

1.1 AI security content in the cyber security training programme
1.2 Ongoing staff awareness of AI threats, with secure AI coding training for developers
2.1 Assess the need, the risks and the mitigations before building an AI system
2.3 Create an audit trail for the AI system
2.4 AI-specific risk assessment and due diligence for external components
2.5 Match intended use to data sensitivity, and encourage staff to report AI security risks
2.6 Least-privilege, risk-assessed permissions for the AI system on other systems
2.7 Due diligence on external providers

Query this from an agent

The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.

MCP endpoint All UK AI Cyber Security Code of Practice controls