Which adventitious sound may clear after the client coughs?

Which of the following best explains why Retrieval-Augmented Generation (RAG) systems require strict access control mechanisms?

Which statement most accurately captures the primary purpose and broader societal significance of explainable artificial intelligence (XAI)?

Which of the following is an advantage of autoregressive AI models for modeling sequential data?

An attacker intentionally modifies training data so that an ML model behaves normally in most situations but produces a specific attacker-controlled output when a particular trigger appears in the input. Which ATLAS tactic does this represent?

Which of the following best reflects the principle of “privacy by design” when developing AI systems?

Which of the following security practices helps detect prompt injection attempts, misuse patterns, and malicious user behavior in operational GenAI systems?

Match each AI/ML attack type with the most appropriate defensive measure. In the mapping below, the numbers represent the attack types, and the letters represent the defensive measures. Select the option that shows the correct mapping. Attack Type 1. Data Poisoning Attacks 2. Model Stealing Attacks 3. Model Inversion Attacks 4. Membership Inference Attacks Defensive Measure A. Regularization and differential privacy B. Data sanitization and anomaly detection C. Rate limiting and output noise D. Data Anonymization

A multinational organization deploys an AI-driven Security Operations Center (SOC) platform that uses behavioral baselining, anomaly detection, and automated containment. After six months, an internal audit reveals that attackers were able to reconstruct sensitive customer information by repeatedly querying the deployed model’s API and analyzing its outputs. No training dataset was directly accessed. Which risk category best describes this attack, and why?

A cybersecurity team needs a generative model for audit-friendly synthetic data generation. They must be able to compute an exact likelihood of each generated sample (for statistical validation), and they prefer a model family that is generally more stable than adversarial training approaches. Which model type best fits these requirements? (select the best answer)