Ethics CT-AI Practice Questions (Page 5)

Questions 21–25

1. 21expert · hard

   A company is developing an AI system for loan approvals. The system is subject to the EU AI Act as a high-risk system. The testing team must ensure that the system meets the requirements. What is the most important testing activity to perform?

   • ATest the system only with synthetic data to avoid privacy issues.
   • BConduct a conformity assessment that includes testing for bias, robustness, and transparency, and document the results.
   • CPerform only functional testing to ensure the system works as intended.
   • DRely on the system's developer documentation to prove compliance.
   Reveal Solution Discussion Select an answer first
2. 22expert · hard

   A company is developing an AI system for hiring. The system is designed to increase diversity, but the test team finds that it inadvertently filters out candidates from certain universities. The company values both diversity and merit. What is the most ethical approach?

   • AInvestigate whether the university filter is a proxy for socioeconomic status and, if so, adjust the model to focus on skills rather than pedigree.
   • BKeep the filter because it is based on objective data about university quality.
   • CRemove the university filter entirely to avoid any bias, even if it reduces the system's predictive accuracy.
   • DAdd a manual review step for candidates from non-target universities.
   Reveal Solution Discussion Select an answer first
4. 23application · medium

   A social media platform uses an AI system to moderate content. The test team is concerned that the system may disproportionately remove posts from minority groups. What is the most important ethical risk to evaluate during testing?

   • ASecurity breach, because the system may be vulnerable to attacks.
   • BPrivacy invasion, because the system analyzes user content.
   • CLack of transparency, because users do not know why their posts are removed.
   • DDiscrimination, because the system may be unfairly targeting certain groups.
   Reveal Solution Discussion Select an answer first
5. 24application · medium

   A multinational company is developing an AI system for hiring. They want to ensure their testing practices are ethically sound and internationally recognized. Which standard would be most appropriate to guide their testing process?

   • AISO 9001, which specifies quality management system requirements.
   • BNIST SP 800-53, which provides security and privacy controls for federal information systems.
   • CIEEE 7000, which provides a process for addressing ethical concerns during system design and testing.
   • DCOBIT, which is a framework for IT governance and management.
   Reveal Solution Discussion Select an answer first
6. 25application · medium

   A team is testing an AI system that recommends criminal sentencing. The testers have access to real historical case data, which includes sensitive personal information. What is the most ethical approach to testing this system?

   • AUse the real data but anonymize it by removing names and addresses.
   • BTest only with data from cases that are already public records.
   • CUse synthetic data that preserves the statistical properties of the original data without exposing real individuals.
   • DSkip testing with real data and rely on the developer's unit tests.
   Reveal Solution Discussion Select an answer first