AI0-001 AI Security Practice Question
A retail company uses a cloud-hosted LLM API to power an internal assistant that answers employee questions about HR policies. The security team discovers that an employee was able to make the assistant output the full text of a confidential severance agreement that exists only in the model provider's training data, not in any company system. Which risk does this incident illustrate?
⚠ Common exam trap
The trap here is labeling any surprising LLM output as prompt injection, when the evidence points to memorized pretraining content rather than attacker-supplied instructions.
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
✓
Training data extraction, where the model memorizes and regurgitates sensitive content from its pretraining corpus.
The assistant surfaced confidential content that only exists inside the provider's training corpus, which demonstrates training data extraction through memorization. This risk is distinct from injection, output handling, and availability threats because the harm is unauthorized disclosure of memorized pretraining data. Organizations relying on third-party models should treat provider training data provenance and memorization behavior as part of their risk assessment.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Insecure output handling, where downstream systems trust model output without validation.
Why it's wrong here
Insecure output handling describes a flaw where an application passes model output into a downstream component, such as a shell or database, without sanitization, leading to code execution or data tampering. In this incident the harm is disclosure of sensitive text directly to a user, not exploitation of a downstream system that consumed the output.
- ✗
Model denial of service, where crafted inputs exhaust compute resources or context windows.
Why it's wrong here
Model denial of service concerns resource exhaustion, such as extremely long prompts or recursive queries that degrade availability or inflate cost. The scenario describes unauthorized disclosure of confidential content, not a loss of availability or resource depletion. Availability attacks leave different evidence, such as timeouts and elevated token consumption.
- ✗
Prompt injection, where an attacker embeds instructions in content the model later processes.
Why it's wrong here
Prompt injection involves malicious instructions hidden in inputs such as retrieved documents or user messages that hijack the model's behavior. Here the model did not follow an injected instruction; it reproduced memorized training text. No evidence indicates an adversarial instruction was embedded anywhere in the interaction chain.
- ✓
Training data extraction, where the model memorizes and regurgitates sensitive content from its pretraining corpus.
Why this is correct
The assistant produced confidential text that exists only in the provider's training data, which is the hallmark of training data extraction. Large language models can memorize rare or repeated sequences and emit them when prompted appropriately. The incident is about memorized pretraining content, not about the company's own systems or prompts being compromised.
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 and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official CompTIA exam blueprint
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.