Automated Regression Testing with Generative AI CT-GENAI Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   When should regression testing be performed in the software development lifecycle?

   • AOnly at the end of the project before release
   • BOnly when the test environment changes
   • COnly when a new feature is added
   • DWhenever code changes are made to the application
   Reveal Solution Discussion Select an answer first
2. 17expert · hard

   A healthcare application requires regression testing with diverse patient data, but using real patient data is prohibited by regulations. The team uses generative AI to synthesize test data. They also need to verify that the synthetic data produces correct outcomes. What is the most effective way to ensure the synthetic data is both realistic and useful for testing?

   • AUse a small set of manually created data that covers only common cases.
   • BGenerate random data and use the AI to create oracles based on the generated data.
   • CUse synthetic data but rely on manual verification of all test outcomes.
   • DGenerate synthetic data that mimics the statistical distribution of real data, and use AI to infer expected outcomes based on the same distribution.
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A financial services company needs to generate realistic test data for regression testing of a new transaction processing module. The data must include a wide variety of customer profiles, transaction amounts, and currency types, but the company has strict data privacy policies that prohibit using real customer data in test environments. Which approach should they take?

   • AUse generative AI to synthesize synthetic data that mimics the statistical properties of real data without containing actual personal information.
   • BUse a small set of manually created test data that covers only the most common transaction types.
   • CRequest permission from customers to use their data in testing, and then use the real data.
   • DAnonymize real customer data by removing names and account numbers, then use it in the test environment.
   Reveal Solution Discussion Select an answer first
5. 19expert · hard

   A company has a large regression suite that takes 6 hours to run. They are introducing a new feature that touches multiple modules. The CI/CD pipeline currently runs the full suite on every merge, causing long feedback times. The team wants to reduce feedback time while maintaining high confidence in the release. They are considering using generative AI to prioritize tests. Which approach best balances speed and confidence?

   • ARun the full suite only on release candidates, and use AI to generate a smaller smoke test for every merge.
   • BIncrease the number of parallel test runners to run the full suite in 2 hours.
   • CUse AI to prioritize tests based on the code changes and historical failure data, running the highest-priority tests first and stopping if a critical failure is found.
   • DUse AI to randomly select 50% of the tests for each merge, alternating the selection.
   Reveal Solution Discussion Select an answer first
6. 20foundation · easy

   Which input is most commonly used by generative AI to create new regression test cases?

   • AThe user manual of the application
   • BExisting test suites and requirements
   • CThe test environment configuration
   • DThe application's source code only
   Reveal Solution Discussion Select an answer first