A structured reference for security architects mapping AI-specific risks to controls from NIST, OWASP, MITRE ATLAS, ENISA, EU AI Act, CISA, Google SAIF, and ISO 42001.