CISA Boosts AI Security With AI Red Teaming & Secure By Design

**Executive Summary:** CISA details enhancements to its Secure by Design initiative by emphasizing AI red teaming and integrating AI Testing, Evaluation, Validation, and Verification (TEVV) into established software testing practices; the document argues AI TEVV should build on traditional software TEVV, describes CISA’s leadership including the TRAINS Taskforce, and promotes post-deployment security testing and collaboration with NIST to ensure safe AI deployment in critical infrastructure.