Supervision of AI Solutions (e.g., Outputs, Impacts, and Decisions) AAIA Practice Questions (Page 2)

Questions 6–10

1. 6foundation · easy

   To whom should supervision findings be reported?

   • AThe general public
   • BOnly the AI model developers
   • CNo one, to avoid unnecessary attention
   • DRelevant stakeholders, including management and compliance teams
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   An e-commerce company uses an AI system to generate product descriptions. The system occasionally produces descriptions with factual errors, such as incorrect product specifications. The company wants to implement a continuous monitoring and feedback mechanism to detect and correct these anomalies. Which mechanism is most effective for this purpose?

   • AImplement automated checks that compare generated descriptions against a database of known product specifications.
   • BRetrain the model weekly with the latest product data to prevent errors from occurring.
   • CRely on customer reviews to identify and report any errors in product descriptions.
   • DConduct a monthly manual review of a random sample of generated descriptions.
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A social media company is developing an AI system to recommend content to users. The company wants to assess the potential impact of the system on user well-being before launch. Which method is most appropriate for this impact assessment?

   • AConduct a user survey to ask users how they think the system will affect their well-being.
   • BRun a controlled experiment with a small group of users, measuring well-being indicators before and after using the system.
   • CConsult with a panel of ethics experts to predict the system's impact on well-being.
   • DReview the system's code to identify any features that could negatively affect well-being.
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   Which role is typically responsible for the ongoing supervision of an AI solution in production?

   • AExternal auditors only
   • BEnd users of the AI system
   • CAI operations team
   • DMarketing department
   Reveal Solution Discussion Select an answer first
6. 10foundation · easy

   What is the first step in handling an error in an AI output?

   • ANotify all customers about the error
   • BImmediately delete the AI model
   • CIdentify and log the error
   • DIgnore the error if it is minor
   Reveal Solution Discussion Select an answer first