AAISM: Advanced in AI Security Management

By Isaca · Question Mode

QuestionQ64

AI Risk Management

An organization is deploying a large language model (LLM) and is concerned that manipulated inputs could compromise its security. Which of the following is the MOST effective way to determine an acceptable risk threshold?

Explanation

An acceptable risk threshold is determined by evaluating the business impact of known threats, so the organization can set risk tolerance according to the consequences of those threats. Logging, input filtering, and output limits are security controls; they do not by themselves define risk appetite or tolerance.

Community Discussion

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