Skip to main content
TPDI logo
Back to Layer 5

Capability 5.3.2

Cybersecurity & Technical Robustness

Definition

Skills to keep AI systems secure, resilient, and compliant with national and international security standards. This includes secure-by-design development, data protection, threat modelling and adversarial testing to guard against attacks and misuse.

Examples

  • ASD/ACSC AI Security Guidance
  • TAFEcyber National Consortium Programs.

Indicators

  • Percentage of AI systems with threat modelling and red-team reports on file
  • Adoption rate of ASD/ACSC 'Engaging with AI' guidance across agencies
  • Incident frequency or severity involving AI components (Cyber Gov reporting)

National AI Plan

PLAN COMMITMENT

Limited commitment

Addresses through broader national cybersecurity and critical infrastructure initiatives and AI-related security considerations (Action 1; Action 7, Action 8), but does not establish or refer to skilling programs to securing AI systems, including secure-by-design development or adversarial testing, hence assessed to be limited commitment. See also 5.3.1 (Assurance & Risk Management). Limited commitment by government is a missed opportunity to build leverage with this existing high agency capability.

Stocktake

Is the Stocktake missing evidence?

Share additional information for potential future Assessments