4.3 Explain the Ethical and Privacy Considerations in AI-assisted Data Analysis, Including Controls to Prevent Data Exposure AI-TECHNICAL-PRACTITIONER Practice Questions (Page 2)

Questions 6–10

1. 6expert · hard

   A global company uses an AI model to personalize marketing emails. The model is trained on customer data from the EU and the US. A data subject in the EU requests access to all personal data the company holds about them, including the data used to train the model. What is the most appropriate response?

   • AProvide only the raw data from the source system, excluding any data that has been transformed for model training.
   • BProvide the raw training data associated with the data subject, including any derived features used by the model.
   • CRefuse the request because the data is used for AI model training and is not subject to GDPR.
   • DAnonymize the data and provide the anonymized version to the data subject.
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A data analyst needs to run a quick exploratory analysis on a dataset containing customer emails and purchase history. The analyst is working on a shared workstation. What is the most appropriate control to prevent data exposure?

   • AAccess the dataset through a secure, authenticated data platform that logs access and enforces least privilege.
   • BEmail the dataset to the analyst's personal email for convenience.
   • CSave the dataset to the shared workstation's desktop for easy access.
   • DUse a personal USB drive to store the dataset and work from it.
   Reveal Solution Discussion Select an answer first
4. 8application · hard

   A social media company uses AI to analyze user posts to recommend content. The company wants to reduce the privacy risk of processing user-generated text. What is the most effective approach?

   • ARequire users to consent to the analysis of their posts.
   • BDelete user posts after they are used for training the model.
   • CUse differential privacy to add noise to the analysis, making it harder to identify individual users in the results.
   • DStore all user posts in an encrypted database.
   Reveal Solution Discussion Select an answer first
5. 9application · hard

   A multinational company processes employee data in the EU and the US. The AI analysis team wants to centralize all data in a US cloud region for model training. The EU data includes personal data of EU residents. What is the primary compliance consideration?

   • AVerify that an appropriate data transfer mechanism (e.g., Standard Contractual Clauses) is in place for EU-to-US transfers.
   • BObtain consent from each employee to transfer their data to the US.
   • CEnsure the US cloud region has strong encryption at rest.
   • DAnonymize the data before transferring it to the US.
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A logistics company uses AI to optimize delivery routes. The AI system analyzes driver locations in real time. The company wants to reduce privacy risks for drivers. What is the most effective approach?

   • ACollect only the location data needed for the current route optimization and delete it after each shift.
   • BObtain consent from drivers to track their locations.
   • CEncrypt all location data in transit and at rest.
   • DAnonymize the location data by removing driver names.
   Reveal Solution Discussion Select an answer first