AI0-001 AI Security Practice Question
A security team is evaluating the risk of adversarial examples against their image classification model. Which characteristic best describes an adversarial example?
⚠ Common exam trap
This exam often tests the distinction between natural misclassifications (due to data quality or model limitations) and intentionally crafted adversarial perturbations, so candidates mistakenly choose options describing data corruption or preprocessing errors instead of recognizing the key element of deliberate, small-scale manipulation.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
An input modified by small, intentional perturbations designed to cause misclassification
An adversarial example is specifically crafted by adding small, often imperceptible perturbations to a legitimate input. These perturbations are designed to exploit the model's decision boundaries, causing it to output an incorrect classification with high confidence. This is a fundamental concept in AI security, highlighting the vulnerability of deep learning models to input manipulation.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
A naturally occurring image that the model misclassifies due to poor training data
Why it's wrong here
An adversarial example is deliberately perturbed to induce misclassification, not a naturally occurring image exposing poor training data. Natural misclassification is tempting because genuine model errors do occur from weak datasets, but that is a data-quality issue, not an adversarial attack.
- ✓
An input modified by small, intentional perturbations designed to cause misclassification
Why this is correct
Small, intentional perturbations exploit the model's learned decision boundaries, shifting a correctly classified image across a boundary without visibly changing it. This satisfies the stem's focus on adversarial risk: the modification is deliberate and imperceptible, distinguishing it from random noise or naturally corrupted inputs, and directly causing misclassification.
- ✗
An image that has been resized incorrectly and appears distorted to the model
Why it's wrong here
Adversarial examples are deliberately crafted perturbations, not incidentally resized or distorted images. Incorrect resizing is tempting because preprocessing errors do cause misclassification, but that is an accidental pipeline fault rather than an intentional evasion designed to fool the model.
- ✗
A corrupted image with missing pixels that the model cannot process
Why it's wrong here
Adversarial examples are intentionally perturbed inputs that remain valid to humans, not corrupted files with missing pixels the model cannot process. Corruption is tempting because it also degrades model output, but it is an availability or input-validity failure, not a crafted evasion.
About these practice questions
One of 962 original AI0-001 practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AI0-001 practice question is part of Courseiva's free CompTIA certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI0-001 exam.