AI Data and Asset Management AAIR Practice Questions (Page 1)

Questions 1–5

1. 1expert · hard

   A company is developing an AI system that uses both public data and proprietary customer data. The customer data is classified as confidential. The team wants to make the training dataset available to external researchers. What is the most appropriate approach?

   • ARemove the confidential customer data and share only the public data.
   • BAnonymize the customer data and share the combined dataset with the researchers.
   • CShare the entire dataset with the researchers, as it is for a good cause.
   • DShare the dataset after signing a data-sharing agreement that restricts use and requires security controls.
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   Which data protection technique involves modifying personal data so that individuals cannot be identified directly or indirectly?

   • AAccess control
   • BAnonymization
   • CTokenization
   • DEncryption
   Reveal Solution Discussion Select an answer first
4. 3expert · hard

   A company is developing an AI model to predict customer churn. They have collected a large dataset containing personal data. The legal team requires that personal data be deleted when it is no longer needed for the purpose. The data science team wants to keep the data for future model improvements. Which approach best balances compliance and business needs?

   • AKeep the data indefinitely for future model improvements, as it is valuable.
   • BDelete the personal data after the model is deployed, but keep the model artifacts.
   • CObtain consent from all customers to keep their data indefinitely.
   • DAnonymize the data after the model is deployed and keep the anonymized dataset for future improvements.
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A retail company is deploying a recommendation system. During a risk assessment, the AI risk manager needs to understand how customer purchase data was transformed and moved from the source system to the model training environment. Which practice would provide the most comprehensive visibility into this data flow?

   • ADocument the data dictionary with field definitions and data types.
   • BCreate a data inventory that lists all data assets and their locations.
   • CMaintain a data lineage record that captures the origin, transformations, and movement of the data.
   • DImplement continuous monitoring of the data pipeline for anomalies.
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   Which data quality dimension ensures that data is current and available within the expected time frame for AI processing?

   • AConsistency
   • BValidity
   • CTimeliness
   • DUniqueness
   Reveal Solution Discussion Select an answer first