QuestionQ30

Securing AI systems

An architect uses the firm’s recommended large language model (LLM) to locate an internal content-management solution. Given the following:

Question Image

Which control is best for mitigating this issue?

Explanation

Response validation applies output guardrails and schema/content checks before an LLM response reaches the user. It can enforce that only the requested application names are returned and block or remove unrelated internal data such as IDs, employee information, risk ratings, and comments.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!