Red Teaming

Adversarial testing where a team simulates attacks against a system to find vulnerabilities before real attackers do. In cybersecurity, this means simulated breaches and social engineering. For AI systems, red teaming involves attempting to elicit harmful outputs, bypass safety filters, or exploit prompt injection vulnerabilities. The EU AI Act and various AI safety frameworks increasingly require or recommend red teaming for high-risk systems.