A financial organization is concerned about the risk of prompt injection attacks on its customer service chatbot. Which of the following controls BEST addresses this concern?

The correct answer is B. Input validation. Input validation inspects and sanitizes user-supplied text before it reaches the model, blocking or neutralizing injected instructions that attempt to override the chatbot's intended behavior. It addresses the attack at the entry point-the earliest and most effective layer of def

AI Security Design and Implementation

Question

Options

AHuman-in-the-loop
BInput validation
CIncreasing model parameters
DContinuous monitoring

How the community answered

(61 responses)

A
2% (1)
B
82% (50)
C
11% (7)
D
5% (3)

Explanation

Input validation inspects and sanitizes user-supplied text before it reaches the model, blocking or neutralizing injected instructions that attempt to override the chatbot's intended behavior. It addresses the attack at the entry point-the earliest and most effective layer of defense for prompt injection. Human-in-the-loop (A) adds oversight but cannot scale to every interaction in a customer service context. Increasing model parameters (C) affects capability, not security against adversarial inputs. Continuous monitoring (D) detects attacks after they occur but does not prevent the injection from executing.

Topics

#Prompt Injection#Input Validation#Chatbot Security#AI Security Controls

Community Discussion

No community discussion yet for this question.

Full AAISM Practice