Responsible Practices and Governance Standards AIE Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   Which technique is commonly used to make AI models more explainable?

   • AEncrypting the model's parameters to protect intellectual property.
   • BIncreasing the number of hidden layers in a neural network.
   • CUsing feature importance scores to show which inputs influenced a prediction.
   • DRemoving all documentation to simplify the system.
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   Which of the following is an example of a continuous improvement process for an AI system?

   • AFreezing the model after initial deployment to ensure stability.
   • BDisabling human oversight to speed up operations.
   • CRegularly retraining the model with new data to maintain accuracy and fairness.
   • DRemoving monitoring tools to reduce complexity.
   Reveal Solution Discussion Select an answer first
4. 13application · medium

   A health tech startup is building an AI model to predict patient readmission risk. They have access to a large dataset of patient records. The startup wants to ensure data governance and privacy compliance. Which practice is most important?

   • AKeep the data on a local server with no internet connection to ensure security.
   • BImplement strict access controls, data encryption, and obtain patient consent for data use.
   • CShare the dataset with external researchers to improve model accuracy.
   • DDe-identify the data by removing names and social security numbers, and assume it is fully anonymous.
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A government agency uses an AI system to flag suspicious financial transactions for investigation. The system's decisions can have serious consequences for individuals. The agency wants to ensure accountability and maintain public trust. Which mechanism is most important to implement?

   • APublish the system's algorithm to the public to ensure transparency.
   • BAutomatically act on all flags to improve efficiency and catch more criminals.
   • CRely on the AI system's high accuracy and assume errors are rare.
   • DEstablish a human review process for all flagged transactions before any action is taken against an individual.
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A European retail company is deploying an AI system that recommends products based on customer purchase history. The system will process personal data of EU residents. The compliance team insists that the system must be able to explain why a specific recommendation was made to a customer, and the data protection officer requires that customers can request deletion of their data used for training. Which combination of actions best addresses these requirements?

   • AAnonymize all training data and disable the deletion feature to avoid conflicts with model performance.
   • BImplement a model-agnostic explainability tool and document the data lifecycle, including retention and deletion procedures.
   • CStore all data in a data lake without access controls to simplify future audits.
   • DUse a black-box model and rely on the customer service team to manually explain recommendations.
   Reveal Solution Discussion Select an answer first