Red teaming
AIGP glossary · Last reviewed: · By Victor Humenhuk (AIGP certified)
Red teaming - Testing security by simulating adversarial attacks to expose flaws before public release.
In the AIGP body of knowledge, Red teaming comes up under Module 7: Governing AI Deployment · BoK III.C; Module 8: AI Governance Vocabulary.
Red teaming in context
- Four definitions recur: ==red teaming== (simulated adversarial attacks), ==challenger models== (raced against the champion on the same data), ==stress tests== (extreme scenarios) and ==threat modelling==. (Periodic assessment - performance, reliability, safety)
Where Red teaming is covered in the AIGP study notes
Related terms
- Challenger model
- Champion model
- Stress test
- Threat modelling
- Automation bias
- Adversarial attack
- Data poisoning
- Data leak
- Deepfakes
- Disinformation
Test yourself on Red teaming
Recognising a definition is not the same as applying it in an exam scenario. Work through the free AIGP practice questions, or read the full AIGP study guide - every study note is free.