MENU
An organization deploying an LLM is concerned input manipulations could compromise security. What is the MOST effective way to determine an acceptable risk threshold?
Suggested Answer: C
Get Premium AAISM Questions
as Interactive Web-Based Practice Test or PDF
Real Exam Format
Detailed Explanations
Instant Access
Get Your Practice Test
AAISM instructs that acceptable risk thresholds must be determined using business impact analysis. This aligns with the broader enterprise risk management principle of defining tolerances based on:
* potential harm
* regulatory exposure
* financial impact
* operational disruption
Monitoring (A) detects attacks but does not set thresholds. Blocking special characters (B) is unrealistic and overly restrictive. Static thresholds (D) ignore business context and practicality.
============================================
17 comments · newest first
VI
Virgina
23 days ago
I lean towards A. Continuous monitoring is key for security.
FR
Frederica
28 days ago
D seems too limiting. We need to allow diverse outputs.
LA
Lashawnda
1 month ago
True, but C helps balance security and functionality.
WI
Willow
1 month ago
I feel like B is too restrictive. We need flexibility in inputs.
ME
Merilyn
1 month ago
A is good too, but C gives a broader view of risks.
VE
Veronika
2 months ago
But what about A? Real-time monitoring could catch issues early.
DO
Dorethea
2 months ago
I agree, C makes sense. It helps prioritize risks.
LE
Leanna
2 months ago
I think option C is the best. Understanding business impact is crucial.
EL
Elvera
2 months ago
D sounds too limiting, we need flexibility!
DI
Diego
2 months ago
A and C together would be ideal, but C is key.
SH
Shantay
2 months ago
Wait, can special characters really be that risky?
AV
Avery
3 months ago
I disagree, A seems more proactive for security.
SH
Shawnda
3 months ago
C is the way to go, gotta know the impact first!
SA
Santos
3 months ago
Implementing a static threshold for LLM outputs could be risky too, since it might not adapt to new threats effectively.
DE
Deja
3 months ago
Restricting inputs with special characters seems too limiting; I feel like it could hinder functionality more than it helps.
ER
Erick
3 months ago
I remember a practice question where we discussed real-time logging and monitoring, but I wonder if that alone is enough to set a risk threshold.
SH
Shasta
4 months ago
I think assessing the business impact of known threats might be the best option, but I'm not entirely sure if it covers all scenarios.