Red teaming
Deliberately probing a model or application with adversarial inputs, including attempts at prompt injection, to surface failures before real users find them. It is a distinct activity from functional testing: functional testing checks the system does what it should, red teaming checks what it should not do but might.
Why exams ask this
Tested as a lifecycle question: when this happens, and who does it. The exam wants adversarial testing before and after launch, by people whose job is to break the system, distinct from ordinary QA that checks the system does what it is supposed to.
Relevant to
AI Fluency: Framework and FoundationsGenerative AI FundamentalsGenerative AI LeaderOCI Generative AI ProfessionalAI Agents CourseGenerative AI and LLMs, NCA-GENLAI Engineering Professional CertificateCertified Machine Learning Engineer, AssociateClaude Code in ActionClaude Certified Architect, Foundations
Related concepts
Resources
No resources linked to this concept yet.