Skip to content

Basic AI Concepts Related to Cybersecurity Questions

Practice questions for Basic AI Concepts Related to Cybersecurity topic in CompTIA SecAI+. 20 questions covering this domain.

20 questions11 easy7 medium2 hard
Q1
medium

A threat actor sends carefully crafted inputs to an AI-based fraud detection model that cause the model to classify fraudulent transactions as legitim...

Q2
easy

Which term describes AI systems that can generate new content — such as text, images, or code — by learning patterns from training data?

Q3
easy

Which AI concept involves training a neural network with multiple layers to automatically extract hierarchical feature representations from raw data s...

Q4
medium

An organization is evaluating AI use cases in its SOC. A security architect wants to implement an AI system that identifies previously unseen attack p...

Q5
easy

Which AI technique involves training a model on a large labeled dataset to predict outputs by learning the relationship between inputs and known corre...

Q6
hard

A SOC team has deployed an AI-based network intrusion detection system. Over three months, the false positive rate has steadily increased while the tr...

Q7
easy

Which term describes a subset of AI that enables systems to learn from data and improve performance over time without being explicitly programmed for ...

Q8
easy

Which AI threat category describes malware that uses AI to continuously modify its own code or behavior to evade signature-based detection?

Q9
easy

What is the primary cybersecurity concern associated with the malicious use of generative AI by threat actors?

Q10
medium

A security analyst observes that an AI-based phishing detection system trained on last year's email data is failing to flag new phishing campaigns tha...

Q11
medium

A security researcher demonstrates that querying an AI model with specific inputs allows them to infer whether specific individuals' medical records w...

Q12
medium

A red team is testing an AI-based spam filter by iteratively modifying phishing emails to find inputs that the model classifies as benign. They use th...

Q13
easy

Which type of machine learning trains an agent to make sequential decisions by rewarding desired behaviors and penalizing undesired ones, rather than ...

Q14
medium

A security architect is reviewing AI risk and notes that the organization's AI threat detection model outputs only a binary label (malicious/benign) w...

Q15
easy

Which term describes a type of attack in which an AI large language model is manipulated into ignoring its safety guidelines by embedding adversarial ...

Q16
medium

An organization is evaluating risks of deploying a large language model that was trained on publicly available internet text. The security team is con...

Q17
easy

Which AI technique is commonly used to generate synthetic but realistic-looking data — such as fake identity documents or synthetic network traffic — ...

Q18
hard

A security team discovers that an adversary has been feeding a continuous stream of subtly mislabeled benign network flows into the feedback mechanism...

Q19
easy

Which AI concept describes the practice of training a single AI model across multiple distributed devices or organizations without any device sending ...

Q20
easy

Which AI concept allows a model trained on a large general dataset to be adapted for a specific security task — such as fine-tuning a language model t...

Sign in to see all 20 questions

Create a free account to browse all questions — completely free during our launch phase.