Free AAIA – Isaca AAIA Practice Test Question

Select an option, then click Submit answer.

• ○ The program includes processes to respond to AI model drift and data integrity attacks.
• ○ The program prioritizes incidents based on alignment with industry leading practices.
• ○ The program uses past AI-related incidents and resolutions to categorize current incidents.
• ○ The program mandates retraining AI systems after incidents are investigated.

Select an option, then click Submit answer.

• ○ Inability to identify where an AI system is housed.
• ○ Difficulty of documenting AI algorithm processes.
• ○ System output not being checked for inconsistencies.
• ○ Cascading failures of AI system outputs.

Select an option, then click Submit answer.

• ○ Underfitting
• ○ Bias
• ○ Overfitting
• ○ Generalization

Select an option, then click Submit answer.

• ○ Filling blank attributes in records with the mean, median, or mode within a grouping
• ○ Placing numerical data into bins or buckets for a manageable quantity of correlations and result analyses
• ○ Separating multiple data attributes within one field into individual attribute columns
• ○ Identifying and deleting duplicate entries in the data set

Select an option, then click Submit answer.

• ○ The space required for AI hardware
• ○ The cost of AI implementation
• ○ The speed of AI implementation
• ○ The ethical implications of AI

Select an option, then click Submit answer.

• ○ Stress tests are regularly conducted to maintain consistent AI performance.
• ○ The applied methodology adequately reflects business objectives.
• ○ Sensitive attributes are converted to other data types prior to input.
• ○ Sampling of AI output is conducted to identify unusual decisions.

Select an option, then click Submit answer.

• ○ Implement model scorecards.
• ○ Clearly define responsibilities for AI risk.
• ○ Have the chief risk officer (CRO) perform a risk analysis.
• ○ Implement human-in-the-loop processes.

Select an option, then click Submit answer.

• ○ Retrain the model immediately and implement privacy-preserving techniques.
• ○ Disable the shared model and notify partners of the potential breach.
• ○ Limit the model's outputs to anonymized results while investigating further.
• ○ Audit the data pipelines of all partners to identify the source of the leak.

Select an option, then click Submit answer.

• ○ Employing specialized tools designed to detect deepfake artifacts and inconsistencies
• ○ Storing all digital evidence on secure remote cloud servers and on local servers
• ○ Requiring digital watermarks for any evidence presented in legal proceedings
• ○ Training legal and forensic professionals to recognize deepfake characteristics in evidence

Select an option, then click Submit answer.

• ○ System algorithms can easily be modified by attackers because the algorithms reside in system memory.
• ○ Project stakeholders may not endorse the system because its behavior may be contrary to their expectations.
• ○ Operational risk may increase because the system is continuously running.
• ○ The system may generate discriminatory output because of biases in training data.

Select an option, then click Submit answer.

• ○ Fine-tune the AI model.
• ○ Scan inputs for code-like structure of text.
• ○ Deploy input validation to sanitize abuse prompts.
• ○ Monitor the prompts for excessive special characters.

Select an option, then click Submit answer.

• ○ It ensures the outputs align with the organization's internal ethical standards.
• ○ It helps auditors identify areas where the AI model may require retraining.
• ○ It allows stakeholders to understand the reasons behind the outputs.
• ○ It reduces the need for human oversight in the loan approval process.

Select an option, then click Submit answer.

• ○ Agent function
• ○ Reward function
• ○ Value function
• ○ Policy function