Question 1

A U.S. government contractor developing an AI threat detection system integrates guidelines that emphasize mapping, measuring, and managing AI risks. Which framework are they applying?

Options :

A : EU AI Act

B : NIST AI Risk Management Framework

C : OECD Responsible AI framework

D : ISO 42001 AI standard

Answer: B

Question 2

Organizations employing MLOps engineers should implement which security measures? (Select TWO)

Options :

A : Immutable audit logs of all pipeline executions and model deployments

B : Providing MLOps engineers with unlimited cloud resources

C : Pipeline version control with mandatory approval for modifications to production workflows

D : Allowing MLOps engineers to directly modify models in production

Answer: A,C

Question 3

A financial services firm notices their fraud detection model's performance degrades after fine-tuning for 50 epochs on new transaction data. Security analysis reveals that attackers injected poisoned samples early in the dataset, and the excessive epochs amplified their impact. What security vulnerability related to epoch configuration does this scenario demonstrate?

Options :

A : Insufficient epochs preventing complete learning

B : Epoch scheduling attacks manipulating learning rate decay

C : Gradient accumulation overflow from too many epochs

D : Overfitting due to excessive epochs amplifying poisoned data patterns

Answer: D

Question 4

A financial institution's AI engine classifies large sets of transaction logs to identify repeated tactics used by known fraudsters, helping analysts recognize emerging fraud methods. Which AI use case is applied?

Options :

A : Incident management

B : Anomaly detection

C : Threat modeling

D : Pattern recognition

Answer: D

Question 5

An autonomous vehicle's AI system passed all safety benchmarks in testing but caused multiple accidents in deployment. Investigation revealed attackers had poisoned the training data with adversarial examples specifically crafted to exploit edge cases not covered in standard safety tests. Which reliability and safety principle would have prevented these failures?

Options :

A : Adversarial robustness testing with continuous safety validation in deployment

B : Testing models on larger datasets

C : Using only simulation data to avoid real-world risks

D : Limiting AI to highway driving only

Answer: A

Higher Test Marks with Free Online CY0-001 Exam Practice

Buying Options: