CompTIA SecAI+ (CY0-001) — Free Practice Questions
10 free sample questions from a bank of 77, with the correct answers and explanations. No signup required — start practising right now.
1Which of the following job roles in an organizational governance structure develops a model from business use cases?
Platform architect
AI risk analyst
Machine learning operations (MLOps) engineer
Data scientist
Answer: D
The short version
Data scientist develops models from business use cases.
Key concepts in this question
organizational governance structure and job roles
business use cases to model development
Platform architect vs AI risk analyst vs MLOps engineer vs Data scientist
Why D is correct
D — Data scientist is the role that builds and trains models from business use cases, selecting features, algorithms, and datasets to meet the defined business problem.
Why the others are wrong
A — Platform architect: designs underlying platform and infrastructure, does not develop the model from use cases.
B — AI risk analyst: assesses governance, bias, compliance, and risk, does not build models.
C — Machine learning operations (MLOps) engineer: deploys, monitors, and operationalizes models, does not develop the model from use cases.
CY0-001 exam tip
If the stem says develops a model from business use cases, choose Data scientist; if it says deploy and monitor, choose MLOps engineer.
2An administrator, who works for a financial institution, is required to implement data security controls for data at rest within AI systems that involve data disclosure. Which of the following is the most suitable control?
Data lineage
Rate limits
Encryption
Masking
Answer: C
The short version
Encryption is the control for data at rest involving data disclosure.
Key concepts in this question
financial institution administrator and data security controls
data at rest within AI systems
data disclosure risk: Data lineage vs Rate limits vs Encryption vs Masking
Why C is correct
C — Encryption protects data at rest by rendering stored AI data, models, and datasets unreadable without keys, directly mitigating disclosure if storage is accessed.
Why the others are wrong
A — Data lineage: tracks origin and flow for governance and traceability, does not protect stored data from disclosure.
B — Rate limits: throttles API and request volume, protects against abuse, not data-at-rest disclosure.
D — Masking: obscures sensitive fields for display, testing, and sharing, not the primary control for data at rest.
CY0-001 exam tip
Data at rest = encryption; data in use/display = masking; abuse/overuse = rate limits.
3A security engineer needs to monitor an AI-based system for runtime operations. The engineer is mostly concerned about the visibility of internal activity. Which of the following is the most appropriate monitoring solution?
Deploying a security information and event management (SIEM) tool
Implementing a web application firewall (WAF) with header logging
Relying on vendor model controls and monitoring prompt inputs
Enabling stack call and debugging level traces at the function level
Answer: D
The short version
Function-level stack call and debug traces give internal runtime visibility.
Key concepts in this question
monitoring AI-based system for runtime operations
visibility of internal activity
SIEM vs WAF with header logging vs vendor model controls and prompt inputs vs stack call and debugging level traces
Why D is correct
D — Enabling stack call and debugging level traces at the function level exposes internal execution, calls, and runtime behavior needed for deep internal-activity visibility.
Why the others are wrong
A — Deploying a security information and event management (SIEM) tool: aggregates external logs and events, does not show internal function-level execution.
B — Implementing a web application firewall (WAF) with header logging: inspects perimeter HTTP traffic and headers, not internal AI runtime activity.
C — Relying on vendor model controls and monitoring prompt inputs: observes only inputs and vendor-provided guardrails, not internal operations.
CY0-001 exam tip
Stem stresses visibility of internal activity at runtime — choose stack/debug traces, not SIEM or WAF.
4Which of the following should an auditor reference when reviewing a company’s human resources AI systems for legal non-compliance?
Organization for Economic Cooperation and Development (OECD) standard
National Institute of Standards and Technology (NIST) AI Risk Management Framework 9RMF)
European Union (EU) AI Act
International Organization for Standardization (ISO)
Answer: C
The short version
Reference the EU AI Act for legal non-compliance of HR AI systems.
Key concepts in this question
auditor reviewing human resources AI systems
legal non-compliance reference
OECD standard vs NIST AI Risk Management Framework vs EU AI Act vs ISO
Why C is correct
C — European Union (EU) AI Act is binding law with risk-based requirements, prohibitions, and obligations for high-risk uses like HR and hiring, so it is the reference for legal non-compliance.
Why the others are wrong
A — Organization for Economic Cooperation and Development (OECD) standard: principles and guidance, not enforceable law.
B — National Institute of Standards and Technology (NIST) AI Risk Management Framework (RMF): voluntary risk-management guidance, not a legal statute.
D — International Organization for Standardization (ISO): voluntary standards, not legislation for non-compliance findings.
CY0-001 exam tip
Legal non-compliance = law — EU AI Act; best practice = NIST AI RMF, OECD, ISO.
5An airline corporation wants to implement a chatbot application using a large language model (LLM) so its customers:
Can ask question and receive answers about flight details.
Have the option to upload files.
Which of the following security controls should the airline use to protect against malicious input and unauthorized use beyond the service-level agreement? (Choose two.)
Prompt guardrails
Role-based access controls
Firewall rules
Model token quotas
Answer: A, D
The short version
Use prompt guardrails plus model token quotas for the LLM chatbot.
Key concepts in this question
airline chatbot using LLM for flight details plus file uploads
malicious input and unauthorized use beyond service-level agreement
Prompt guardrails vs Role-based access controls vs Firewall rules vs Model token quotas
Why A and D are correct
A — Prompt guardrails filter and constrain malicious input, jailbreaks, and out-of-scope prompts. D — Model token quotas limit consumption and unauthorized use beyond the service-level agreement.
Why the others are wrong
B — Role-based access controls: restricts who can access functions by role, does not filter malicious prompt input or enforce token/SLA usage.
C — Firewall rules: controls network ports and traffic, does not inspect LLM prompts or meter tokens.
CY0-001 exam tip
For LLM input attacks choose guardrails; for overuse/cost/SLA abuse choose token quotas — this stem needs both.
6A security operations center (SOC) has a very high volume of logs and alerts. The manager proposes the implementation of machine learning (ML) system to help with triage. Which of the following tasks is most suitable?
Applying filters on specific alerts
Automatically patching vulnerable systems
Identifying and classifying alerts
Summarizing the content of alerts
Answer: C
The short version
Use ML triage to identify and classify high-volume alerts.
Key concepts in this question
security operations center (SOC) with very high volume of logs and alerts
machine learning (ML) system for triage
Applying filters vs Automatically patching vs Identifying and classifying alerts vs Summarizing content
Why C is correct
C — Identifying and classifying alerts is the ML triage fit: clustering, prioritization, and categorization to route high-volume alerts efficiently.
Why the others are wrong
A — Applying filters on specific alerts: static rule suppression, does not require ML triage.
B — Automatically patching vulnerable systems: remediation action, not alert triage.
D — Summarizing the content of alerts: generative summarization helps readability but is not the core triage task of classification and prioritization.
CY0-001 exam tip
SOC overload + ML triage = identify and classify; patching and filtering are not ML triage.
7An organization recently created a custom model that integrates with a language model (LLM). The developer notices that the application programming interface (API) costs have increased. Which of the following is the best control to reduce cost?
Implementing prompt templates
Increasing central processing unit (CPU) and memory
Reducing the model size
Adjusting token limits
Answer: D
The short version
Adjust token limits to reduce increased API costs.
Key concepts in this question
custom model integrating with LLM and increased API costs
best cost control: prompt templates vs CPU/memory vs model size vs token limits
Why D is correct
D — Adjusting token limits directly caps input/output tokens billed per API call, reducing cost without redesigning the model.
Why the others are wrong
A — Implementing prompt templates: standardizes prompts and quality, does not directly cap billed usage.
B — Increasing central processing unit (CPU) and memory: adds infrastructure cost and does not lower API token charges.
C — Reducing the model size: major re-architecture with retraining and quality impact, not the best immediate cost control.
CY0-001 exam tip
API cost spike = token limits first; infrastructure scaling does not fix token billing.
8A security administrator needs to improve an AI model. During an initial investigation, the administrator notices that two successive login features are recorded every day, and then a successful login occurs after a specific time interval. All the successful login attempts have been during office hours.
Which of the following techniques should the administrator use to improve the AI model’s security?
Access management
Pattern recognition
Signature matching
Vulnerability analysis
Answer: B
The short version
Use pattern recognition for repeated login timing and office-hours behavior.
Key concepts in this question
improve AI model security from login observations
two successive login features recorded every day, then successful login after specific interval, during office hours
Access management vs Pattern recognition vs Signature matching vs Vulnerability analysis
Why B is correct
B — Pattern recognition learns the repeating sequence — two daily features plus time-interval success in office hours — to detect deviation and improve the model.
Why the others are wrong
A — Access management: enforces permissions and access decisions, does not learn temporal login patterns.
C — Signature matching: matches known malicious signatures, not behavioral sequence learning.
D — Vulnerability analysis: finds weaknesses in systems, does not model repeated login behavior.
CY0-001 exam tip
Repeated interval and time-of-day behavior in the stem signals pattern recognition, not signatures.
9Which of the following is the most concerning risk for a company that allows corporate end users to use public-facing large language models (LLMs)?
Inaccuracies due to hallucinations
Out-of-date acceptable use policies
Data security regulatory violations
Malicious code generation
Answer: C
The short version
Public LLMs risk data security regulatory violations from user inputs.
Key concepts in this question
corporate end users using public-facing LLMs
most concerning risk: hallucinations vs acceptable use policies vs regulatory violations vs malicious code
Why C is correct
C — Data security regulatory violations is most concerning because sensitive corporate data pasted into public models can leak, breach disclosure duties, and trigger compliance penalties.
Why the others are wrong
A — Inaccuracies due to hallucinations: quality risk, less severe than regulated data exposure.
B — Out-of-date acceptable use policies: governance gap that enables risk, not the resulting most concerning risk itself.
D — Malicious code generation: possible misuse, but data exfiltration and regulatory violation is the primary enterprise concern.
CY0-001 exam tip
Public LLM + end users = data leakage and regulatory violation as the top answer over hallucinations.
10Which of the following requires developers to harden infrastructure to protect AI systems?
Intake processes
Acceptable use policies
Development guidelines
Configuration standards
Answer: D
The short version
Configuration standards require developers to harden AI infrastructure.
Key concepts in this question
developers harden infrastructure to protect AI systems
Intake processes vs Acceptable use policies vs Development guidelines vs Configuration standards
Why D is correct
D — Configuration standards define required hardened baselines and settings developers must apply to infrastructure hosting AI systems.
Why the others are wrong
A — Intake processes: governs request and approval for new systems, does not specify hardening settings.
B — Acceptable use policies: defines allowed user behavior, does not mandate infrastructure hardening.
C — Development guidelines: advises secure coding practices, not infrastructure hardening baselines.