Which of the following BEST describes an adversarial attack on an AI model?

The correct answer is B. Providing inputs that mislead the AI model into incorrect predictions. An adversarial attack is specifically defined as crafting inputs - often with subtle, imperceptible perturbations - that cause an AI model to make incorrect predictions or classifications. Option B is the textbook definition. Option A (attacking hardware) is a…

AI Security Design and Implementation

Question

Options

AAttacking the underlying hardware of the AI system
BProviding inputs that mislead the AI model into incorrect predictions
CReverse engineering the AI model using social engineering techniques
DConducting denial-of-service (DoS) attacks against AI APIs

How the community answered

(56 responses)

A
2% (1)
B
88% (49)
C
4% (2)
D
7% (4)

Explanation

An adversarial attack is specifically defined as crafting inputs - often with subtle, imperceptible perturbations - that cause an AI model to make incorrect predictions or classifications. Option B is the textbook definition. Option A (attacking hardware) is a physical/infrastructure attack, not an AI-specific adversarial attack. Option C conflates reverse engineering with social engineering and doesn't accurately describe adversarial ML. Option D (DoS against APIs) is an availability attack targeting infrastructure, not the model's decision-making logic.

Topics

#Adversarial attacks#AI model security#Model integrity#Input manipulation

Community Discussion

No community discussion yet for this question.

Full AAISM Practice