Fundamentals of testing AI-based systems · Question 47 of 60
A team testing a virtual assistant cannot agree on "correct" answers because users have very different expectations. Which cause of the test oracle problem is present, and what helps?
2 answers are correct. Hard
Show answer
Answer
- A The subjectivity of behavior
- B Helpful are, for example, domain experts, agreed output boundaries or specialized techniques such as A/B testing
Explanation
The correctness of an AI system's behavior can be subjective, for example because expectations of virtual assistants vary widely. Approaches include domain experts, agreed output boundaries and specialized techniques such as A/B testing, which compare behaviors without needing an expected output for every case.