Testing generative AI · Question 54 of 60
A company wants to use red teaming to check an LLM with RAG integration for security vulnerabilities before launch. Which attacks should be considered?
3 answers are correct. Hard
Show answer
Answer
- C Indirect prompt injection
- D Non-AI security factors
- E Malicious content hidden in the documents used by RAG
Explanation
When red teaming is about security, you check how vulnerable the system is from outside, with classic as well as AI-specific weaknesses. The latter include indirect prompt injection and documents with hidden instructions that get into the system via RAG.