Preparing Practice Exam…
Please wait a moment while the system generates the exam questions, answers, and study session.
Please wait a moment while the system generates the exam questions, answers, and study session.
ISACA Advanced in AI Audit (AAIA) · Isaca
Select one answer.
Written by other learners. Comments can disagree with the marked answer — check the official documentation when in doubt.
Data sanitization ✅ BEST (removes risk) Data masking ⚠️ Partial protection
As per ChatGPT and Gemini - Winner.Data sanitization is the most effective and primary control to prevent sensitive information disclosure by LLM chatbots because: It ensures sensitive data is removed or cleaned before it ever reaches the model. Prevents leakage at the source level, not just during output. Works for: Training data Prompt inputs Knowledge bases (RAG systems) Eliminates risks like: PII exposure Confidential data leakage Prompt injection exploiting sensitive data
C, from 1.14, data masking and tokenization can help