Questions
Common questions.
What is AI Red Teaming?
AI red teaming is the practice of systematically probing an AI system for vulnerabilities, biases, and failure modes by simulating real-world attacks and edge cases.
How does AI Red Teaming work?
It is increasingly required by regulation for high-risk AI systems.
Which terms are related to AI Red Teaming?
Closely related concepts include MITRE ATLAS, Prompt Fuzzing, AI Risk Assessment, DORA (Digital Operational Resilience Act). Each is defined in the Prefactor glossary.
Keep reading