Europe’s AI Security Standard: Vulnerability Disclosure + Red Teaming

### Executive summary ETSI EN 304 223 establishes baseline cybersecurity requirements for deployed AI systems across their lifecycle, stressing continuous vulnerability disclosure, mandatory security and adversarial testing, and recurrent AI red teaming; the report explains these requirements and recommends practical actions—publish disclosure policies, focus testing on exploitability and impact, define retest triggers, and engage independent testers—to help organizations prepare for European expectations.